working paper 07-2 offshoring, outsourcing, and production relocation--labor-market effects in
Post on 03-Feb-2022
4 Views
Preview:
TRANSCRIPT
Working Paper S e r i e s
W P 0 7 - 2 A P r i l 2 0 0 7
P E T E R G . P E T E R S O N I N S T I T U T E F O R I N T E R N A T I O N A L E C O N O M I C S
1 7 5 0 M A S S A C h U S E T T S A v E N U E , N W | W A S h I N G T O N , D C 2 0 0 3 6 - 1 9 0 3
T E L : ( 2 0 2 ) 3 2 8 - 9 0 0 0 | F A x : ( 2 0 2 ) 6 5 9 - 3 2 2 5 | W W W . P E T E R S O N I N S T I T U T E . O R G
Offshoring, Outsourcing, and Production Relocation—Labor-Market Effects in the OECD Countries and Developing AsiaJacob Funk Kirkegaard
AbstractThis working paper evaluates the validity of available data on and the extent of the impact of offshoring on service-sector labor markets in the United States, EU-15, and Japan. A three-tier data validity hierarchy is identified. The impact of offshoring on employment in the three regions is found to be limited. Correspondingly, developing Asia is unlikely to experience large employment gains as a destination region. The paper highlights the case of the Indian IT industry, where the majority of job creation has been in local Indian companies rather than foreign multinationals. Domestic entrepreneurs have played a crucial role in the growth of the Indian IT-related service industry. However, increased tradability of services and associated skill bias in favor of higher skilled workers could have an uneven employment impact on developing Asia. Some high-skilled groups are benefiting and will continue to benefit dramatically from new employment opportunities and rising wage levels. Meanwhile, the same skill bias may eliminate many employment opportunities for unskilled or low-skilled groups in the region. Developing Asian countries therefore face a double educational challenge in the coming years: the need to simultaneously improve both primary and higher education.
JEL codes: ���,���, J�1Keywords: Service Sectors, Offshoring, Production Relocation, Data Source Validity,
Automation, Highly Skilled Workers
Jacob Funk Kirkegaard is a research associate at the Peterson Institute for International Economics. His current research focuses on European economies and structural reforms, pension systems, offshoring, and the impact of information technologies. Before joining the Institute, he worked with the Danish Ministry of Defense and the United Nations in Iraq. In 1996 he graduated from the Danish Army’s Special School of Intelligence and Linguistics with the rank of first lieutenant. He is coauthor of Transforming the European Economy (�004). This working paper is a revised version of the paper first presented at the �006 SCAPE Workshop “Economic Implications of Globalization, Production �ragmentation and Outsourcing for Singapore and Asia.” It is forthcoming in the Singapore Economic Review. The author would like to thank Professors Ramkishen S. Rajan and Shandre M. Thangavelu for many helpful comments. Responsibility for any errors rests solely with the author.
Copyright © 2007 by the Peterson Institute for International Economics. All rights reserved. No part of this working paper may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying,
recording, or by information storage or retrieval system, without permission from the Institute.
�
“Outsourcing is a particular type of international trade. We are used to trade in goods, but trade in services has expanded recently, made possible in large part by advances in telecommunications. Like all forms of international trade, outsourcing benefits an economy overall, though there are also short-term costs as workers are displaced.” Greg Mankiw, Chairman, Council of Economic Advisers1
“India is a labour-oriented country, so we need more jobs that are secure…. [O]utsourcing does not lead to more efficiency. It creates complications in rendering good customer services.”
R. J. Sridharan, General Secretary, All India Bank Officers’ Association�
�ew new topics in international economics have risen faster to the top of the political agenda, while si-
multaneously so poorly understood and quantified, than outsourcing in recent years. Also termed offshor-
ing, international sourcing, production fragmentation, delocalization, production relocation, and even
global supply chain management, this phenomenon’s alleged hugely disruptive effects on employment
have caused considerable concern among large groups of people and, by extension, politicians. Unfortu-
nately, the near complete absence of valid empirical evidence on the phenomenon has allowed entrepre-
neurs and consultants to frame the debate to promote their own interests, thus fueling public anxiety.
This working paper surveys the scattered empirical evidence gathered to date in the Organization for
Economic Cooperation and Development (OECD) countries and attempts to draw from that evidence
a number of conclusions regarding the present and future scope of the phenomenon’s impact on employ-
ment. It gauges which economic actors could gain or lose in both OECD and developing countries. The
empirical evidence in this working paper focuses on “production relocation” rather than the broader
category of foreign direct investment (�DI) in an economy.
Section I provides the necessary terminology and presents the most important empirical findings
to date in the United States, EU-15, and Japan. Section II surveys different estimates for employment
potentially affected by offshoring and offshore outsourcing, while presenting estimates from detailed US
information technology (IT) occupations indicating very extensive skill bias. The potential impact of
automation on the service sectors is also evaluated. Section III considers the impact on developing Asia
of the findings in sections I and II and presents data from India pointing to the domestic origin of most
employment growth. Section IV concludes the working paper.
1. Comment in response to a question on “Ask the White House,” December 5, �00�. See Mankiw and Swagel (�005) for an exposé on US politics in this subject.
�. Comment in “India Bank Workers Protest Outsourcing,” Financial Times, July �8, �006.
�
I. TERMINOLOGY AND EMPIRICAL FINDINGS
Terminology
A good starting point in the search for a framework for discussion of the phenomenon is the matrix
developed by the UN Conference on Trade and Development (UNCTAD), among others, in its World
Investment Report �004. This matrix is based on the production location and ownership options facing a
company (figure 1).�
Companies reorganizing their original nucleus of internal, domestic assets and deciding to relocate
a captive unit for intermediate inputs to a foreign country subsequently offshore production. Compa-
nies choosing instead to rely on a domestic but external nonaffiliated producer outsource production,
and those choosing to rely on an external nonaffiliated producer in a foreign country offshore outsource
production.
The focus here is “production relocation”—i.e., instances of (1) total or partial closure of an enter-
prise’s existing production units in one country with accompanying work force reductions, and (�) either
the opening of affiliates abroad for the production of the same goods or services (offshoring) or the forg-
ing of a subcontracting contract with a nonaffiliated firm for production of the same goods or services in
their home country (outsourcing) and/or abroad (offshore outsourcing).
OECD (�005a, section 5.4.�) makes a further useful distinction between two types of subcontract-
ing.4 The first one is commonplace “low-tech subcontracting” in areas such as catering, cleaning, trans-
portation, or standard design of “nuts and bolts” intermediate inputs for various kinds of machinery. The
second category concerns areas and items with far higher technological content, such as particular research
studies, accounting, research and development projects, advertising, customer relations, or IT services.
Here the supplying firm will not be delivering a standard good, but rather it must adhere rigorously to
the primary manufacturer’s technical specifications. This working paper focuses exclusively on the second
category of subcontracting, which IT has in recent years increasingly made tradable across international
borders, especially as a novelty also incorporating many low-wage countries.
It is important to note that production relocation forms only a (small) subset of the long-term in-
crease in global �DI and expansion of the activities of multinational companies (UNCTAD �004, �005;
OECD �006a). The empirical magnitude of any impact of the broader rise in �DI on home-country
�. This working paper uses the terminology in this matrix.
4. It must be stressed that this section OECD Handbook concerns “areas of future investigations” and therefore does not represent a consensus view of definitions or terminology.
�
employment—i.e., of the “jobs forgone and never created at home in the first place because investments
are located elsewhere” type—is outside the immediate scope of this working paper.5
The standard matrix presented in figure 1, and the accompanying terminology, suffers from theoreti-
cal incompleteness in the description of production location choices, in that it describes only the traffic
on one side of the street so to speak—the arrows point only “outward,” away from what is as a rule a
high-wage country of origin. Implicitly, this matrix frames companies’ production relocation choices in a
globalizing world economy as a one-way movement—typically from a high-wage location to a low-wage
destination.6
This incompleteness seems almost certain to have had a distortionary effect on the political discus-
sion of production relocation in high-wage countries. Also, survey methodologies of empirical studies of
offshoring and offshore outsourcing recurrently reflect this incompleteness. These studies typically cover
only companies’ decisions to relocate production abroad, while explicitly excluding measures of compa-
nies choosing to relocate production to, for instance, the United States or an EU-15 country—in other
words, engage in (viewed typically from an OECD country) “insourcing/inshoring” of tasks and jobs.
Several studies have attempted to focus on the “two-way street” of international engagement in another
way in the United States, by highlighting the many jobs created in the United States by foreign investors.
Many of these, however, are greenfield projects or results of mergers and acquisitions and therefore fall
outside the narrower definition of insourcing/inshoring adopted here (see, for instance, Slaughter �004
and Zeile �005).
�igure � presents a more complete matrix of the production relocation options facing a multina-
tional company. Unfortunately, this matrix is almost wholly theoretical in nature, in the sense that no
major empirical study has so far attempted to implement this methodology in its entirety for production
relocations. Very few studies (Jensen, Kirkegaard, and Laugesen �006) come close, making empirical “net
employment effect” estimations of companies’ production relocation decisions extremely rare.
The reasons for the near complete empirical neglect “of the other side of the street”—inshoring and in-
sourcing—are several. One is that the media (and by extension probably politicians) tends to focus almost
exclusively on the “bad news of job losses,” while ignoring the “good news of job creation.” Another is
that surveys of companies capturing the entire figure � matrix would have to be (1) very extensive, raising
5. Desai, �oley, and Hines (�004, �005) establish that �DI by US multinational corporations (MNCs) is complemented by increases in the domestic capital stock of these MNCs, while confirming earlier findings in �eldstein (1995) that broader measures of �DI (i.e., including first-time �DI by non-MNCs) reduce domestic investments.
6. To a degree, this reflects a theoretical inability to cope with firm heterogeneity. The matrix in figure 1 is implicitly designed to describe the process, which occurs when a successful hitherto domestic-only OECD country–located small or medium-sized company expands to multinational status. This standard model does not capture the additional options available to companies that have already achieved multinational status, nor can it immediately describe the choices available to companies expanding internationally from a non–OECD country home base.
�
concerns of respondent burden, and (�) carried out over a very large population in order to cover a signifi-
cant number of companies engaging in either (or both) types of sourcing and hence would be very costly.
Empirical Findings to Date
With public anxiety about jobs running high and in the absence of valid official statistical data� on the ex-
tent of the impact on employment of offshoring and offshore outsourcing, in particular, two broad types
of proxy data sources have emerged:8 “private consulting type estimates” and “press monitoring estimates.”
Both, however, possess inherent flaws.
“Private consulting type estimates”9 are obtained typically from small population surveys of a partic-
ular private company’s clients’ previous actions and future intentions with respect to production location
and ownership. Hence, these surveys are rarely based on samples that can reasonably be labeled represen-
tative, and the accompanying estimates and projections are frequently obtained with less than transparent
methodologies.
�urthermore, the actual sample data and answers are almost always confidential and can hence not
be queried or independently verified. This is particularly unfortunate given the obvious incentives for
many private providers of such surveys—who also frequently advise their clients on how to implement
production location strategies—to establish offshoring and offshore outsourcing as “the next big thing.”
A comprehensive methodological critique and argument concerning (potential) “selection bias” in
many of these “private consulting type estimates” is thus problematic, as they generally do not provide any
information about their methodologies and data collection strategies. Indeed, this omission is itself the
most important disqualifying factor. An influential example is �orrester Research, Inc. (�004a), which is
an update on a frequently quoted �00� �orrester Research note. It predicted that �.� million US service
jobs will have moved offshore by �015. The study actually provides a “methodology section” (reproduced
in box 1).
�. A �004 report from the US Government Accountability Office on the availability of official US data on the subject had the telling title, “Current Government Data Provide Limited Insight into Offshoring of Services” (GAO �004). Recently the MIT Industrial Performance Center Working Group on Services Offshoring reached the same result in its final report, stating as its second conclusion that “The data currently available for characterizing and measuring services offshoring have severe limitations” (Industrial Performance Center �006, iv). See also GAO (�005a).
8. Domestic outsourcing, which encompasses the geographic shifting of jobs within a country with an immediately assumed first-order zero net effect on employment, is implicitly captured in official estimates of gross job gains and losses. See, for instance, the US Bureau of Labor Statistics Business Employment Dynamics Program at www.bls.gov/bdm/home.htm. It is not obvious that national authorities would have any use for or interest in gathering data on the employment impact of domestic outsourcing beyond this type of data, except when it is an automatic, if peripheral, creation arising from other data gathering activities.
9. See, for instance, �arrell and Rosenfeld (�005), �orrester Research, Inc. (�004a, �004b), Boston Consulting Group (�004), Deloitte Research (�004), Roland Berger and UNCTAD (�004), and Global Insight (�004).
�
A detailed discussion of the relationship between �orrester’s results and the Bureau of Labor Statis-
tics data can be found in Kirkegaard (�004), but suffice to say here that questions concerning (potential)
selection bias when making economywide predictions could be raised with each individually listed input:
Who are the 100+ companies interviewed? Are they large or small? Are they multinationals? Who are the
1,800+ industry participants surveyed? How is “vendor briefing” defined? Who inquired about offshoring
and concerning what? Who are the third-party sources?
As such, studies of this type by and large represent the lowest level in the validity hierarchy of em-
pirical data. Yet, this does not indicate that they cannot be useful, as they will frequently be the only data
available—only that the user should proceed with caution.
“Press monitoring estimates”10 are a systematic gathering of press reports concerning the offshoring
or offshore outsourcing of jobs from a country. While this approach obviously depends crucially on the
questionable assumption of complete press coverage, it does have the advantage that it relies on broadly
verifiable public sources.
Two slightly different methods can be identified: scanning news media for reports either of all types
of corporate restructuring or only of offshoring and offshore outsourcing related to corporate restructur-
ing. The former method has the advantage of allowing comparisons of the relative importance of different
causes of corporate restructuring—i.e., seeing if, for instance, people are being laid off due to corporate
bankruptcies rather than offshoring and offshore outsourcing.11
10. See, for instance, Bronfenbrenner and Luce (�004), E�ILWC (�004), Gerstenberger and Roehrl (�006), and Kirkegaard (�005).
11. If corporate bankruptcies and aggregate demand were the major problems, clearly the policy response to rising unemploy-
Box 1 Forrester’s methodology
Inputs:
Four trips to India 100+ user company best practices interviews 1,800+ surveys of North American businesses and IT leaders �00 vendor briefings �00 inquiries on offshoring Research with third-party sources
Steps in the update process:
Update the baseline numbers by job categories based on �00� Bureau of Labor Statistics data Reassess the four offshore growth curves and add year-by-year granularity for �00�–08 Reevaluate the offshoreability of the different job categories based on: 1) the nature of the work: Is it knowledge-intensive versus transaction-focused and to what degree is it process automated; �) the skills available offshore; �) offshore investments to support the process by third-party IT and business process outsourcing service providers. Internal peer review
�
On the other hand, it is clearly very resource demanding and time consuming to systematically cover
all types of corporate restructuring, and therefore the threshold for inclusion into the “sample of
restructurings” will frequently be put so high, say at 50 or 100 layoffs, that much valuable information
on smaller restructurings goes unrecorded. Such thresholds may dangerously bias the sample toward “big
events,” such as bankruptcies or plant closures (internal restructurings), which will tend to pass a high
threshold in terms of layoffs, whereas smaller-scale events, say offshoring of part of the 50-employee com-
pany back-office, may not.
Therefore the “hard number” of layoffs caused by offshoring and offshore outsourcing from this
broad type of media survey will likely be an underestimate. Yet, these studies—if sufficiently funded and
carried out by independent organizations—should be viewed as the “middle rung” in the validity hierar-
chy of empirical data.
No systematic collection of official data by national or international statistical entities specifically
targeting the employment impact of offshoring and offshore outsourcing, which would have constituted
the top (i.e., broadly acceptable for academic work) rung of the data validity ladder, exists anywhere today.
The closest in existence is “derived official data” from other current official data collection programs.
US Empirical Data
The US Bureau of Labor Statistics (BLS), as part of the Mass Layoff Statistics (MLS) Program, has col-
lected such “derived official data” since �004. The MLS program collects from administrative unemploy-
ment records and employer interviews data regarding the reasons for production establishment closings in
the United States. As in most media monitoring estimates, the threshold for inclusion is layoffs of 50 or
more employees.1� BLS data can therefore be expected to suffer from a similar downward bias with regard
to the actual number of layoffs.
Since �004, the BLS has published data on the number of jobs lost in mass layoffs associated with
both domestic and overseas relocations of production by US businesses. The total for the entire two-year
period from �004Q1 to �005Q4 is presented in table 1.
Just under a million Americans lost their jobs as a result of unscheduled (i.e., excluding seasonal
work and vacation periods)1� mass layoffs, and approximately 1� percent of these layoffs can be attributed
ment would include other measures than if it were mostly related to offshoring and offshore outsourcing.
1�. �or more information on the program, see the BLS MLS Web site at www.bls.gov/mls/home.htm. Also, as noted in Schultze (�004), the MLS datasets cannot, like this working paper, take account of “jobs forgone in America” via reduced hiring.
1�. The inclusion of “vacation period” as a reason for mass layoff may seem a non sequitur but is related to the fact that the collection of MLS data is based on individual unemployment records, and a person may be eligible for unemployment benefits during the period a given plant is shut down for an extended holiday period.
8
to “movement of work”—i.e., domestic in-house relocation, outsourcing, offshoring, and offshore out-
sourcing. This 1� percent number is relatively large, although not as big a reason for mass layoffs as,
for instance, “contract completion” (~�5 percent) or downsizing (~16 percent) and roughly on par with
layoffs due to “bankruptcy” and “financial difficulty” combined.14
It is possible, based on available data, to calculate the relative importance of the four types of “move-
ment of work.” However, given that detailed information regarding only �� percent of the total num-
ber of relevant separations is available (line 4), estimating a precise “hard number of jobs lost” to either
category of “movement of work” (lines 5 to 8) is not appropriate. Instead, the far right column of table 1
estimates the share of total separations associated with movement of work attributable to each category.15
Line 5 shows that more than half (55 percent) of all separations associated with “movement of work”
is of the domestic in-house type and therefore have neither ownership nor international locational effects.
They are therefore outside the scope of this working paper.
Line 6 shows that the apparent employment impact of offshoring as a result of domestic outsourc-
ing only is very limited: only 1.4 percent of all separations. This number seems very low and possibly
indicates that in most instances, domestic outsourcing has already occurred or is carried out without any
large-scale loss of jobs. In lines � and 8, we see that the combined employment effects of offshoring and
offshore outsourcing represent just 4 percent of all separations from mass layoffs in the United States in
�004–05, with in-house offshoring representing three-quarters.
The data presented in table 1 should obviously not be taken as the “definitive truth” on the employ-
ment effects of offshoring and offshore outsourcing, given the significant methodological shortcomings
previously described. Yet, with the broadest and most valid available empirical evidence largely free of se-
lection bias showing that just 1 in �5 US separations in mass layoffs can be attributed to either offshoring
and offshore outsourcing, one cannot escape the conclusion that the heated public and political debate on
the issue has been vastly overblown at least in the United States and that the direct employment effects are
very limited. This is especially the case in the US economy as a whole, where very large numbers of jobs
are constantly created and destroyed. The latest available data for �005Q� show that more than 8 million
jobs were created, while �.4 million were lost.16
However, it is important to note that this conclusion, reached on the basis of even the US BLS
data—which is at the top of the data validity ladder—is a “snapshot,” i.e., merely states what the status
14. �or data on these and other reasons for mass layoffs, see detailed BLS MLS data available at http://data.bls.gov.
15. It is assumed in this estimate that the approximately �8 percent of separations, for which no detailed information is available, is distributed similarly to the �� percent for which information is available. It is possible that employers not wishing to reveal their engagement in, for instance, offshoring actively try to hide this fact from BLS interviewers. This would cause the estimate in lines 6, �, and 8 in table 1 to be downward biased by an unknown amount. However, was such bias to exist it would be extremely unlikely to be of a magnitude that would materially impact the conclusions drawn below.
16. See data for gross job gains and losses at the BLS Business Employment Dynamics Program at www.bls.gov/bdm/home.htm.
�
is at a given point in time. It does not consider any dynamic employment effects arising out of even this
limited level of offshoring and offshore outsourcing. The almost inevitable “technology spillovers” from
offshored or offshore outsourced production to foreign domestic companies may lead to increased com-
petitive pressure in the US (or other rich-country) home market as well.1� Such dynamic spillover effects
may lead to adverse employment effects at competing home-country firms, if they as a result of increased
foreign competition are forced to downsize or even go into bankruptcy. Due to the still very limited dura-
tion of time-series, none of the empirical data on offshore outsourcing presented in this working paper
can reasonably be utilized to estimate these effects.
European Empirical Data
Turning to the empirical evidence in Europe, one is unfortunately forced to take a step down in the data
validity hierarchy and instead of relying on “derived official data” make do with data from a “press moni-
toring estimate.”
Since �00�, the European Commission has via the European Restructuring Monitor (ERM) financed the
systematic monitoring of European newspapers and business press18 for instances of corporate restructur-
ings in all sectors of the EU economies leading to job losses. The threshold for inclusion is 100 job losses
or at least 10 percent of the work force at companies employing more than �50 people. This means that,
as was the case with the US BLS data above, these data are also likely to suffer from omission bias related
to the threshold of inclusion.19
The most worrying additional bias in these data comes from the difference in the size of EU coun-
tries: Press coverage of restructuring events above the threshold in smaller countries seems more certain
than in larger ones. It thus seems highly likely that a 100+ layoff would be reported in Luxembourg but
less likely in the German press. �or purposes of direct comparison with US data, only the results from
�004–05 are shown in table �.
As with the US data, it is essentially meaningless to come up with a “hard number” for EU-15
job losses, given the methodological weaknesses of the data.�0 Instead table � estimates the relative impor-
1�. Such spillovers in the service sectors seem most likely to occur as a result of the rapid employee turn-over in the sector seen, for instance, in the Indian offshore centers recently. Unlike in the manufacturing sector, where foreign technology blueprints and physical production facilities will frequently be relatively expensive for domestic competitors to copy (although of course nowhere nearly as expensive as developing them themselves), in many service sectors, when employees leave a foreign company, they may take with them “inside their heads” so to speak, many of the most important competitive parameters. These will be intangible things such as “knowledge of” the rich world’s export market, multinational-company managerial techniques, or potential rich-world client base.
18. See the European Restructuring Monitor at www.emcc.eurofound.eu.int/erm for a complete list of covered media.
19. The potential downward bias from the inclusion threshold is even higher in these data than in the US BLS data. �or a detailed description of included data, see the European Restructuring Monitor at www.emcc.eurofound.eu.int/erm.
�0. Aubert and Sillard (�005), based on a different methodology utilizing data for individual firms and certain assumptions
10
tance of different reasons for industrial restructurings in the EU-15. Two things are immediately clear:
�irst, at only a combined 5 percent of total covered job losses in the EU-15, job losses due to domestic
outsourcing, offshoring, and offshore outsourcing are dwarfed in importance by job losses caused by cor-
porate bankruptcies and internal restructurings (downsizings). Second, this limited employment impact is
roughly equivalent to that recorded in the United States.�1
It seems appropriate to state that the employment effects of offshoring and offshore outsourcing in
the EU-15—like in the United States—have also been vastly exaggerated, as they are very limited.
The EU-15 data from the ERM offer two additional levels of relevant empirical detail, namely sector
and country specificity. A breakdown of EU-15 job losses from offshoring and offshore outsourcing in
�004–05 reveals that manufacturing,�� which made up only 18 percent of all total employment during
�004–05,�� still dominates job losses, accounting for 56 percent of all job losses. However, service sec-
tors—financial, information and communication technologies (ICTs), and others—accounted for the
remaining 44 percent of job losses�4—i.e., sectors traditionally thought to be less affected or unaffected
by international trade do account for almost half the job losses. This topic is revisited in section II.
A breakdown of the EU-15 data into individual country data for job losses from offshoring and offshore
outsourcing presents a reasonably valid capture of cross-country differences because the ERM data are col-
lected in a similar way across all EU-15 countries.
�igure � shows an offshoring/offshore outsourcing intensity index for the individual EU-15 coun-
tries for �004–05. The index is estimated as a ratio of the number of job losses in each country and total
private employment in the country with the EU-15 average for the period equal to 100. It is evident in
figure � that the intensity of offshoring and offshore outsourcing varies significantly within the EU-15,
ranging from close to four times the average in �inland to effectively no impact in Greece and Luxem-
bourg.
The detailed causes of these wide intra-EU-15 differences are outside the scope of this working pa-
per, but in brief it seems that local, country-specific factors dominate. In the case of �inland, for instance,
regarding relocations abroad, estimates that approximately 95,000 manufacturing-sector jobs were lost in �rance alone between 1995 and �001.
�1. Domestic in-house production relocation—in the ERM data in table � labeled relocation—is significantly smaller in the EU data at 1.5 percent versus 6.5 percent of covered job losses. However, given the much larger geographic size of the US domestic economy relative to individual EU economies, this is not surprising. Van Welsum and Vickery (�006) estimates based on a multivariate regression model that, for all OECD countries, only a very small number of terminations are related to offshoring and offshore outsourcing.
��. In the ERM data, manufacturing is defined as the sum of the chemical, construction and woodwork, electrical goods, food, beverages and tobacco, glass and cement, metal and machinery, motor, pulp and paper, and textile and leather sectors.
��. EU-15 manufacturing employment data are from the Eurostat Labor �orce Survey at http://epp.eurostat.ec.europa.eu.
�4. In the ERM data, the ICT sector is defined as the sum of the information technology and postal and telecommunication sectors. The financial services sector itself and “others” is a residual category.
11
the presence of low-wage and recent EU member-state Estonia, which has close cultural and linguistic ties
to �inland, as a neighboring country can partly explain this country’s high intensity score.
There does not seem to be a strong correlation between the intensity of offshoring and offshore
outsourcing in individual EU-15 countries and the structural rigidities of their labor markets. �igure 4
(left axis) plots country offshoring and offshore outsourcing intensity levels and the score from the OECD
for that country’s employment protection legislation (EPL) concerning collective dismissals of the kind
typically covered by ERM data. As can be seen from the near horizontal grey trend line in figure 4, there
is essentially no link between the two variables.
�igure 4 (left axis) indicates that even countries with very heavy EPL against collective dismissals
have not escaped the employment effects of offshoring and offshore outsourcing. Hence, attempts to
protect jobs from offshoring and offshore outsourcing by making large layoffs increasingly difficult for
employers seem to have very limited effect.
This apparent lack of a direct relationship may, however, mask differing dynamic labor-market de-
velopments over the course of the business cycle between countries with a very high EPL level and those
with a lower level. �or instance, in periods of cyclical expansion, companies in countries with very restric-
tive EPL can be expected to rely on offshoring or offshore outsourcing to a larger degree than those in
countries with less restrictive EPL.�5 Companies in high-EPL countries can be expected to take advantage
of offshoring in order to avoid the risk of being stuck with “unsackable workers” in the event of a cyclical
downturn. This means that domestic cyclical labor-market swings as a result of offshoring and offshore
outsourcing may become more pronounced in countries with lower levels of EPL, as domestic hiring oc-
curs more readily. At the same time, more jobs will likely be permanently forgone in high EPL countries,
which will be prone to chronically high levels of unemployment instead of cyclical labor-market swings.
Combining the cross-country variance in offshoring and offshore outsourcing intensity in the EU-15
with a measure of public anxiety over this issue from the EU-wide Eurobarometer public poll,�6 it can be
shown that there is no statistical relation between the intensity of offshoring and offshore outsourcing in a
country and its public’s concerns regarding the phenomenon. This is illustrated on the right axis of
figure 4.
The lack of a relationship between the “facts on the ground” and “public opinion” regarding off-
shoring and offshore outsourcing further underlines that the origin of the debate in the EU-15 is purely
political.
�5. The German Chamber of Commerce—Germany is a restrictive EPL country—explicitly lists this as a major reason for the attraction of German companies toward offshoring to Eastern Europe. See “Working Longer to Save Jobs: But Will It Help Europe Close the Productivity Gap?” Financial Times, July ��, �004.
�6. The data is from Eurobarometer 6� from the spring of �005. Like the ERM data, the Eurobarometer has the advantage of being collected in a uniform manner across all EU member states. See details of the program and data at the European Commission Web site, http://ec.europa.eu.
1�
�inally it should be mentioned that at least one large-scale empirical study (Jensen, Kirkegaard,
and Laugesen �006) exists for an EU-15 country (Denmark) that incorporates both offshoring/offshore
outsourcing and inshoring/insourcing decisions, making a “net employment effect” estimate possible. This
study indicates that for a small, open economy with a labor market significantly more flexible than its
neighboring countries in the European Union, the net employment effects of companies’ production re-
location decisions are indeed positive. It shows that offshoring and offshore outsourcing of predominantly
low-skilled jobs in the manufacturing and IT sectors are more than offset by inshoring of higher skilled
jobs in both the manufacturing and service sectors.
Japanese Empirical Data
Empirical data of the type utilized for the United States and the European Union do not seem to exist for
Japan. Instead, this author finds only indirect evidence from data on broader categories of �DI and global
production integration.
Japanese manufacturing companies have for decades been expanding their participation in global
and particularly East Asian production networks concentrating in the machinery and transportation
equipment sectors. Kimura (�005) cites several Japanese studies for the finding that outward Japanese
�DI has had a small but negative effect on domestic manufacturing employment. �igure 5 seems to su-
perficially confirm this finding. It shows domestic Japanese manufacturing employment and employment
in overseas affiliates of Japanese manufacturing companies. �rom �001Q� to �006Q1, domestic Japanese
manufacturing employment declined by 1.� million to 11.5 million, while employment at overseas manu-
facturing affiliates increased by 900,000 to �.8 million. �igure 6 breaks down overseas affiliate employ-
ment by region.
The fact that employment at Chinese manufacturing affiliates accounts for more than 60 percent of
the total increase and by �005 surpassed ASEAN-4–located employment clearly indicates the rapid rise of
China as the most important destination for labor-intensive Japanese �DI in the manufacturing sector.
Yet, data in figures 5 and 6 give only a tentative indication of the fact that production relocation
via offshoring and offshore outsourcing in Japan is likely occurring or possibly accelerating. The decline
in domestic manufacturing employment may have occurred predominantly due to accelerating labor
productivity growth in the Japanese manufacturing sector. Japan and Iceland were the only two OECD
countries where productivity growth (measured as value added per person employed) in the manufactur-
ing sector was higher in 1995–�00� than in the preceding period 1990–95. �rom 1995 to �00�, Japanese
manufacturing-sector productivity growth at just above 4 percent surpassed that of the United States.��
��. Due to very low productivity growth in the service sectors, overall Japanese productivity growth during this period, however, was significantly below that of the United States. Data from OECD (�005b).
1�
Meanwhile, the rapid expansion of affiliate employment may be due to a need to serve new markets
in, for instance, China from local production facilities. Such an interpretation is supported by the rela-
tive stability in the share of sales by Japanese overseas manufacturing affiliates going back to Japan—at
about 9 percent from �001 to �006. Local sales make up an overwhelming �0+ percent of total affiliate
sales, although showing a slight relative decline, offset by a relative increase in sales by affiliates to third
countries.�8 In both cases any attribution of the observed domestic employment developments to either
offshoring or offshore outsourcing would be erroneous.
A Ministry of Economy, Trade and Industry survey (METI �006, �9) indicates the potential relative
resilience of domestic Japanese manufacturing employment to offshoring and offshore outsourcing. The
survey, which covers fiscal �00�, shows that the vast majority of Japanese manufacturing companies main-
tain their domestic production activities at the previous level following an expansion overseas. Sixty-four
percent of the respondents (4,440 Japanese manufacturing companies) state that “overseas production has
not affected domestic employment and production,” while a further 19 percent reports that “overseas pro-
duction has shifted domestic employment and production towards highly value-added products.” In total
more than 80 percent of Japanese domestic manufacturing employment thus seems unaffected by com-
pany expansion overseas. Only 6 percent of companies report that domestic plants have been closed and
employment reduced, another 6 percent has decreased domestic production and reduced employment but
avoided plant closures, while the remaining 5 percent of companies have decreased domestic production
without direct job losses.
While systematic bias in the answers to this survey cannot be ruled out, as Japanese manufacturing
companies answering the survey may have had an interest in hiding possible domestic layoffs, the survey
does suggest the possible presence of complementarities in domestic and overseas investments by Japanese
companies, similar to what Desai, �oley, and Hines (�004, �005) found for US companies.
No employment data are available for Japanese �DI outside the manufacturing sector, although
METI, based on a study of domestic employment growth and balance of payment statistics, in its White
Paper on International Economy and Trade �004 concluded that “[i]n the context of the services industry
as a whole, both production value and employment figures position the professional and business services
industry in Japan as an industry at the growth stage…. Japan’s offshore outsourcing remains insignificant”
(METI �004, chapter 1).
In summary, data on the employment effects of offshoring and offshore outsourcing in OECD
countries remain scarce. The little data that have been gathered from the most reliable derived official data
and media monitoring exercises indicate limited direct employment effects of offshoring and offshore out-
sourcing to date: only 4 to 5 percent of total large-scale layoffs in the United States and EU-15. However,
�8. Data are from the METI Survey of Trends in Business Activities of Japanese �oreign Affiliates (Kaiji Chosa).
1�
just because the employment effects of offshoring and offshore outsourcing have hitherto been limited
does not mean that they are not novel or would not have a potentially huge impact. The next section ad-
dresses this topic.
II. HOW BIG CAN IT GET IN SERVICES?
Production relocation, offshoring, and offshore outsourcing of jobs in the manufacturing sector have been
occurring for decades (see, for instance, Lewis and Richardson �001, �eenstra 1998, and Zeile �00�).
However, with rapid technological development, investments in ICT infrastructure, accompanying
declines in the costs of global communication, ongoing trade liberalization in many service sectors (via
gradual implementation of the General Agreement on Trade in Services [GATS] schedule as well as uni-
lateral and bilateral liberalizations), and the emergence of new low-wage suppliers of skilled labor (most
noticeably India), international sourcing of IT and ICT-enabled business services—such as customer rela-
tions services, back-office services, and numerous professional services—has risen rapidly. Estimates of the
industry’s growth vary substantially, but one indicator is the rise in global revenues of the Indian software
and IT-enabled service sector. According to the Indian IT industry association NASSCOM, these rev-
enues have risen at a compound annual growth rate of �8 percent since 1999 to more than $�8 billion in
�005 (NASSCOM �006a).
This section will provide—in the words of Van Welsum and Vickery (�005)—an overview of the
“outer limits” of the potential employment impact of this new development in the nonmanufacturing,
service sectors, while illustrating the emergence of deepening skill bias in IT employment, and contem-
plate the effects of possible automation of service jobs.
Gauging “Potentially Affected Employment”
In earlier periods when offshoring and offshore outsourcing were almost exclusively occurring in the
manufacturing sector, it was straightforward to quantify the potential employment effects of the trend by
viewing the share of employment in the manufacturing sector as the “outer limit.” While any prediction
of almost complete disappearance of employment (though not necessarily output) in the manufacturing
sector in any OECD country is of course extreme, this approach did make considerable short-term politi-
cal sense in that laid off manufacturing workers could be encouraged to seek employment in the allegedly
unaffected (and much larger in employment terms) service sectors. However, as services have become
more tradable�9 and offshoring and offshore outsourcing have spread to the service sectors (see section I
�9. �or excellent empirical descriptions of this rise in tradability of services, see annual data for US trade in services available at www.bea.gov.
1�
above showing 44 percent of layoffs in the EU-15 are in the service sectors), such a simple sectoral demar-
cation of the potential employment effects is no longer tenable.
Instead careful consideration must be given to the precise impact and intensity of the use of IT
throughout the highly heterogeneous service sectors. Rather than using a sectoral demarcation as before,
identifying the most affected occupations—i.e., specific job categories across sectors—is now the most
fruitful approach. However, as the vast majority of statistical employment data are collected on a sectoral
basis, this means that once again researchers face significant data constraints.
With the study of the occupational impact of offshoring and offshore outsourcing still in its infancy,
no consensus view has so far emerged regarding the choice of criteria by which “occupations affected by
offshoring and offshore outsourcing” should be identified. Several, sometimes overlapping, suggestions
have been made. Van Welsum and Vickery (�005) list the following criteria:
• people holding jobs where they are likely to make intensive use of ICTs to produce output;
• their output can be traded/transmitted with the help of ICTs (ICT-enabled trade in services);
• the work has a high explicit information or “codified knowledge” content (and no or little tacit or
implicit knowledge); and
• the work does not necessarily require face-to-face contact.
In addition, the authors, inspired by Bardhan and Kroll (�00�), list the following potentially im-
portant criteria, not explicitly taken into account: (1) a high-wage differential with similar occupations in
destination countries; (�) low set-up barriers; and (�) low social networking requirements.
Based on these criteria, the authors estimate that in �00� the sum of identified potentially affected
occupations made up 19.� percent of total employment in the EU-15, 18.6 percent in Canada, 19.4
percent in Australia, 18.1 percent in the United States (�00� data), and 1� percent in South Korea (�00�
data).�0 In all cases, except South Korea, this share is larger than the total share of employment in the
manufacturing sector.�1 In other words, in the affected occupations, offshoring and offshore outsourcing
potentially have a very large effect on employment.
Other studies using different methodologies have reached different but still large estimates. Bardhan
and Kroll (�00�) calculate 11 percent of US employment in occupations potentially affected by offshor-
ing in �001. Garner (�004) estimates 10 percent of total US employment in potentially affected occu-
�0. Van Welsum and Vickery (�006) note that Korea is in the process of revising its occupational classification system, which is expected to have a material impact on the estimated share of potentially affected employment.
�1. Manufacturing-sector employment in �00� was as follows: EU-15, 18.6 percent; Canada, 14.6 percent; Australia, 11.4 percent; United States, 1�.� percent (�00�); and South Korea, 19.1 percent (�00�). Source: OECD Labor Statistics and Eurostat Labor �orce Survey.
1�
pations in �000. �arrell and Rosenfeld (�005) estimate 11 percent of US service-sector employment is
theoretically affected but only � percent actually will be affected by �008. Blinder (�005, �006) adopts a
hybrid sectoral-occupational approach and focuses on the split between personally and impersonally deliv-
ered services and estimates that up to �1 percent of US employment in �004 may be potential candidates
for offshoring or offshore outsourcing. Jensen and Kletzer (�005) adopt a spatial clustering approach,
where industries and occupations highly clustered within the United States are assumed to be interna-
tionally tradable. They estimate that 1�.5 percent of US employment is in jobs where both industry and
occupation are tradable.
Levy and Murnane (�004, �006a, �006b) emphasize that for a job to be offshored or offshore out-
sourced, it (1) could be done anywhere and (�) must concern information that can be exchanged between
the client and the offshore producer without misunderstandings—i.e., must be largely rule-based. They
(wisely perhaps) do not offer numerical estimates but contend that the majority of jobs relocating in the
future will be such rule-based middle-skilled jobs.
In short, most estimates of employment potentially affected by offshoring and offshore outsourcing
indicate that the share is larger in the service sectors than in the manufacturing sectors.
“Affected Jobs” Do Not Equal “Jobs Lost” to Offshoring and Offshore Outsourcing
Technological advances especially in the ICT sector have, as noted, been instrumental in increasing
the tradability of services in recent years.�� However, several studies indicate that technological change
concerning computers also may have another, far bigger and disruptive impact on labor demand. Autor,
Katz, and Krueger (1998), Autor, Levy, and Murnane (�00�), and �alk and Koebel (�004) with data from
the United States and Germany, respectively, show that increases in computer capital in nonmanufactur-
ing industries are closely related to non–trade related declines in the demand for low-skilled workers and
routine tasks.
This dual role and simultaneous effect of IT, facilitating both the increased tradability of services
as well as a non–trade related increase in the destruction of low-skilled jobs through automation, raises
the risk that offshoring and offshore outsourcing will be blamed for disruptions in OECD-country labor
markets far above their actual impact, not least due to the difficulty—keeping in mind the absence of
the targeted empirical data collection illustrated in section I—of disentangling these dual effects. Hence
the risk that “potentially affected jobs” come to represent “jobs lost/to be lost to offshoring and offshore
outsourcing.”
��. This section builds extensively on the analysis found in Mann and Kirkegaard (�006).
1�
An illustration of this issue can be found in the recent employment developments in US IT-related
occupations, all of which in the studies listed above are included in the group of “potentially affected oc-
cupations.”��
Table � looks at the developments in net employment in low- and high-wage US IT-related occupa-
tions from 1999�4 to mid-�005, and as can be seen the trends are dramatically different. Low-wage (a rea-
sonable proxy for low-skilled) IT-related employment in occupations earning on average significantly less
than the US average wage, in areas such as call centers, data entry, and word processor typing, declined
by more than a third, or by more than �50,000 jobs. Yet, at the same time employment in high-wage (a
reasonable proxy for high-skilled) IT-related jobs in occupations earning about twice the US average wage,
in areas such as software engineering and network administration and analysis, increased by 18 percent, or
almost 400,000 jobs.�5
Total net employment in the selected US IT-related occupations being down by more than �50,000
over the period masks two highly diverging employment trends in affected IT-related occupations, namely
that (1) low-wage/skilled US IT jobs are disappearing faster than even blue-collar US manufacturing-sec-
tor jobs;�6 and (�) high-wage/skilled US IT-related jobs continue to be created at a pace five to six times
faster than the growth of the total US work force.��
This finding leads to two additional themes. �irst, it is not immediately possible to estimate how
much of the decline in low-wage/skilled employment is due to offshoring and offshore outsourcing and
how much is due to technological innovation itself. However, keeping in mind the findings in section I,
but also factors such as the very high overall productivity growth in the US service sectors�8 in the period
and its links to IT capital deepening,�9 it seems at least probable that the latter dominates. Hence, many
��. Table � for illustrative and recognition purposes uses detailed 4-digit BLS Occupational Employment Statistics (OES) data, which are significantly more detailed than any of the listed studies. However, in all relevant cases, the higher-level less-detailed occupational categories to which those listed in table � belong are included in the “potentially affected categories.”
�4. Some degree of arbitrariness exists in the selection of a starting year for this type of calculation. By choosing 1999, a deliberate attempt is made to “slice the peak of the Internet boom off” and achieve a starting point approximating “sustainable employment levels.” US total employment by the end of the March–November �001 recession was back at the approximate level of late 1999. Total US nonfarm employment in December �001 equaled 1�0.� million, while in December 1999 it was 1�0.5 million. Source: US BLS Current Employment Statistics (CES) data.
�5. US average—not to be mistaken for the largely stagnant median—wages increased from $��,800 in 1999 to $��,900 in �004.
�6. The much-publicized decline in the number of production workers in the US manufacturing sector from 1999 to May �005 was “only” �4.� percent (US BLS CES data).
��. It grew only approximately �.5 percent over the period (US BLS CES data).
�8. Van Ark (�005) shows that US productivity (measured as GDP per hour worked) in ICT-using services from 1995 to �00� rose by an average annual 5.� percent, as opposed to only 1.� percent in the EU-15, accounting for almost the entire difference in overall productivity growth between the two areas.
�9. See, among others, Jorgenson and Stiroh (�001).
18
more low-skilled IT-related jobs in the United States seem to have been destroyed by technological inno-
vation than lost to offshoring and offshore outsourcing.
Second, table � shows that it is quite possible for high-wage/skilled job categories to be widely re-
garded as “affected by offshoring and offshore outsourcing,” while maintaining rapid employment growth.
Yet, looking at specific categories of high-wage/skilled US IT-related employment, computer program-
mers40 have experienced significant net declines of more than a quarter in employment. Meanwhile,
higher-wage/skilled software engineers41 have experienced large increases of more than 50 percent in
employment, so that indeed there has never been more software engineers employed in the United States
than now (mid-�005 data).
It is subsequently possible to speculate that at least some of this increase in employment in the high-
est-wage/skilled US IT-related occupations may be a result of inshoring and offshore insourcing (see figure
�) of this type of highly specialized work to the United States from other countries. Ironically, such a sce-
nario would create a positive effect on employment in these highest-wage/skilled occupations from being
potentially affected by offshoring and offshore outsourcing. The prospect of overall improved employment
opportunities in occupations affected by offshoring and offshore outsourcing is also found in Amiti and
Wei (�005, �006), who use different data.
According to table �, substantial skill bias in employment is affecting ICT-intensive occupations
potentially affected by offshoring and offshore outsourcing. Job destruction is rapid among lower-wage/
skilled and even select high-wage/skilled jobs, while job creation continues to be brisk among the high-
est-wage/skilled jobs. This does not mean, however, that even the highest-wage/skilled jobs are entirely
unaffected by the potential for offshoring and offshore outsourcing. Some evidence suggests that such
occupations, which in terms of employment opportunities were effectively immune to the business cycle,
may now have become far more cyclical than during earlier business cycles (illustrated in figure �).
The figure shows how the unemployment rates of US IT-related occupations during the 1991 reces-
sion never went above 4 percent and until �000 were generally less than half the overall US unemploy-
ment and close to essentially full employment (allowing for a small amount of transitional unemploy-
40. Job classification is defined in the OES as: 15-10�1 Computer Programmers: Convert project specifications and statements of problems and procedures to detailed logical flow charts for coding into computer language. Develop and write computer programs to store, locate, and retrieve specific documents, data, and information. May program web sites.
41. The two job classifications for software engineers are defined as: (1) 15-10�1 Computer Software Engineers, Applications: Develop, create, and modify general computer applications software or specialized utility programs. Analyze user needs and develop software solutions. Design software or customize software for client use with the aim of optimizing operational efficiency. May analyze and design databases within an application area, working individually or coordinating database development as part of a team. Exclude Computer Hardware Engineers (1�-�061). (�) 15-10�� Computer Software Engineers, Systems Software: Research, design, develop, and test operating systems-level software, compilers, and network distribution software for medical, industrial, military, communications, aerospace, business, scientific, and general computing applications. Set operational specifications and formulate and analyze software requirements. Apply principles and techniques of computer science, engineering, and mathematical analysis. Source: US BLS OES Program, www.bls.gov/oes.
1�
ment). However, during the �001 US recession, unemployment in US IT-related occupations not only
rose to unprecedented levels but actually rose above the overall unemployment level. Subsequently, it
declined dramatically in �004–05 to again reach its historical level of full employment at approximately �
percent.
How much of the spike and subsequent rapid decline in unemployment in IT-related occupations
is due to increased tradability and offshoring and offshore outsourcing, and how much is due to a large
number of probably ultimately insufficiently qualified people entering these occupations during the IT
boom of the late 1990s and exiting after the bust, a number that cannot be immediately determined? �ig-
ure � does, however, suggest that US IT-related occupations, even among the highest-wage/skilled, which
were to a large extent unaffected by unemployment may now be facing a far more “normal job market,”
where short-term spells of unemployment cannot be ruled out during recession periods.4�
This interpretation of the data—that the labor market for IT-related occupations is starting to
behave more like a regular cyclical labor market, at least partly as a result of the increased tradability of
IT-related services—corresponds with the description in Bhagwati, Panagariya, and Srinivasan (�004),
among others, of offshoring and offshore outsourcing as simply an expansion of trade into new areas.
Indeed many high-skilled IT-related workers in the United States have since �000 faced a “new situa-
tion” of possible short-term unemployment. However, this group should probably be viewed as entering
a situation of “normal labor-market circumstances” rather than one where employment opportunities are
unavoidably and irreversibly disappearing.
Increased Tradability as a Step Toward Automation?
The dual role of IT as facilitating both tradability of services and destruction of employment of low-
skilled workers points to a further future impact on service-sector employment—that of automation. Levy
and Murnane (�006) conclude that “At least for the moment, there is a strong overlap between the jobs
threatened by globalization and the jobs threatened by computerized work.” Agreeing with this conclu-
sion, one can ask how long this overlap of jobs being offshored or offshore outsourced and automated will
last before automation completely takes over.
As Levy and Murnane (�004, �006) point out, one of the most important things that makes service-
sector tasks and jobs increasingly tradable and possible to offshore and offshore outsource—namely that
they can be described in relatively simple rules—is also the very thing that frequently makes it possible for
a computer to substitute for a human in processing information and performing the task/job. Therefore
4�. This also partly reflects the fact that a far larger share of people in US IT occupations today work in “normal non-IT related” sectors of the economy. Mann and Kirkegaard (�006) present estimates that upto two-thirds of the US IT occupation work force are today employed in non-IT producing sectors.
�0
it may be appropriate to start viewing offshoring and offshore outsourcing as merely a temporary middle
stage before many of these tasks are completely automated. Even the lowest-cost human labor cannot
realistically compete against the lowest-cost labor of all, namely computers themselves.
Research by IBM and Datamonitor4� indicates that average costs per customer interaction (for
instance, a call to a company representative) costs approximately $6, if the receiving company employee is
located in a high-wage country. Offshoring such services (for instance, call centers) to India reduces costs
by up to 50 percent to just about $� per interaction. However, relying instead on email-based communi-
cation saves a further perhaps 10 percent on costs, relative to the costs of a rich-world agent. Yet, utilizing
even more IT to replace humans and implementing interactive voice response services, automated speech
recognition, or Web-based self service lowers costs by up to 90 percent.
Evidently, while there may be very large labor cost savings related to offshoring and offshore out-
sourcing from a rich country to, for instance, India, these savings are again dwarfed by the savings avail-
able to companies relying wholly or partly on computers to carry out the interaction. With the ongoing
rapid development of speech-recognition technology, it seems probable that an increasing number of
companies will switch to automated computer-handled customer relations,44 making the offshoring and
offshore outsourcing of such tasks merely a midpoint.
This does not mean that employment in the destination countries for offshored and offshore-out-
sourced services is facing any immediate large-scale impact from automation of service tasks, but it does
indicate that the required skill threshold for possessing a “secure job” in high-wage countries has seem-
ingly been driven higher in recent years by both offshoring and offshore outsourcing and IT itself. So will
technology in the future also gradually raise the skill threshold for the type of jobs that cannot be auto-
mated in the service sectors and therefore are potential profitable candidates for offshoring and offshore
outsourcing? Offshore destinations for services, too, will therefore constantly need to improve the skill
level of their workers and cannot hope to rely chiefly on expanding low-wage work forces as their prin-
cipal appeal. This may in the long term have some employment impact on developing Asian countries, a
topic that will be developed in section III.
In summary, section II has shown that occupations “potentially affected” by offshoring and offshore
outsourcing make up very large shares of OECD-country employment. Yet, rather than a uniform decline
4�. IBM estimates partly derived from author conversations with IBM representatives and also referenced in the IT Review: Voice Recognition Section, Financial Times, May 4, �005, and “Whatever Happened to Customer Service?” USA Today, September �6, �00�. Datamonitor estimates cost savings from offshoring to India at �0 percent and from automation at 80 percent. See also Karen Bannan, “Ernestine, Meet Julie,” CFO Magazine, January 1, �005.
44. See, for instance, “Technology Quarterly,” The Economist, December 6, �001, and Katie Hafner, “A Voice With Personality—Just Trying To Help,” New York Times, September 9, �004. This author can testify to this development in his own retail bank (Citibank) and wireless phone company (Cingular), where in �00� depending on the time of day his call was received by a representative in either the United States or India. Today, he reaches a voice-recognition menu irrespective of the time of day.
�1
in such employment, affected work forces in rich countries are experiencing a rapid movement away
from low-wage/skilled jobs toward higher-wage/skilled employment. Data from the United States suggest
that in recent years this movement has happened faster in IT-related occupations than among blue-collar
workers in the manufacturing sector. IT has facilitated this shift toward higher-wage/skilled employment,
through both increased tradability of services and technological innovation independently. The same IT
may in future lead to increased automation of some service-sector tasks, highlighting the need for skill
upgrading among work forces in offshoring and offshore outsourcing destinations.
III. IMPACT ON DEVELOPING ASIA
The impact on developing Asia of the topics covered in sections I and II comes in three main areas: (1)
direct employment impact of production relocation and offshoring and offshore outsourcing; (�) in-
creased tradability of services and accompanying skill bias; and (�) the possible premature emptying of the
appropriately priced pool of talent in developing Asian countries.
Direct Employment Impact
Section 1 revealed that production relocation and offshoring and offshore outsourcing have had only a
relatively minor impact on OECD-country labor markets. The flip side of the coin is that potential des-
tination countries in developing Asia have not seen a huge inflow of such jobs. Even allowing for signifi-
cantly more labor-intensive production in recipient countries (see below and �arrell and Rosenfeld �005),
the number of jobs terminated in origin countries is so relatively low that the magnitude of jobs received
is nondramatic seen from the perspective of recipient countries. The International Labor Organization
(ILO �006, ��) notes that in �00� the entire work force of the Indian IT industry—of which, as will be
shown below, far from all work at multinational companies—accounted for only two-tenth of 1 percent
of the country’s labor force.
�urthermore, as almost the entire cost savings from offshoring and offshore outsourcing decisions
typically come from wage differentials, there is an upper limit to just how much more labor intensive
low-wage destination-country production can become, relative to home-country production, while still
representing a profitable business decision. And with a larger number of staff in an initially low-wage
destination country, companies will be even more sensitive to rises in wage costs.
In short, entrepôt economies aside, in all likelihood developing countries in Asia will not get a broad
and significant employment boost from offshoring or offshore outsourcing per se. Even India, by far the
most successful destination for offshored and offshore-outsourced IT service jobs, as well as the economy
��
best positioned to reap by far the largest future direct employment gains from offshoring and offshore
outsourcing, has not achieved such a boost.45
According to NASSCOM (�006a),46 India’s IT industry trade body, industry employment grew
from �84,000 in fiscal 1999–�000 to 1,�8�,000 in fiscal �005–06 at the rate of roughly �00,000 people
per year.4� Some of this employment growth would have been for servicing the small, but growing, do-
mestic Indian IT market, but a large majority would have occurred in the export-oriented sector. NASS-
COM reports that in fiscal �005 exports accounted for $1�.� billion of the total IT services and software
revenue of $��.6 billion, or almost 80 percent.48
More importantly, though, revenue statistics for the Indian IT service sector (figure 8) show that the
largest share of employment growth in the Indian IT service export sector is a result of the highly success-
ful growth of indigenous Indian IT companies—such as Tata Consulting Systems, Infosys, Wipro, and
Satyam to name the top four—rather than the relocation of production to India by foreign multination-
als.
Of course, such ownership demarcation between domestic and foreign owners in a highly global-
ized and dynamic industry is only a snapshot: India-located IT companies are constantly bought and
sold across borders. �or instance, in November �004 GE sold its Indian technology service subsidiary,
GECIS,49 while in April �004 IBM bought the Indian call-center and back-office service provider Daksh
eServices for $150 million.50 Keeping this caveat in mind, it can be seen in figure 8 that Indian companies
dominate the largest IT-enabled service segment, while foreign multinationals51 are larger in the smaller
business process outsourcing (BPO) segment of IT services. Assuming that local Indian IT service com-
45. See, for instance, A. T. Kearney (�004) for this estimate. �or a perhaps slightly biased view, see NASCCOM-McKinsey (�005).
46. NASSCOM data are clearly at the lowest level of the data validity hierarchy described in section I.
4�. �iscal year equals the period from April to March.
48. Some uncertainty surrounds NASSCOM’s export data for Indian IT services and software, which are, for instance, frequently up to 10 times as high as the corresponding US import data. See GAO (�005b) for elaboration on this issue. This discrepancy is rooted in the fact that NASSCOM collects data from members in a form that is not completely compatible with the IM� balance of payment data collection standards. Due to the on-site (in export markets) delivery model (GATS mode 4) of many Indian companies, many “export revenue” figures reported by NASSCOM will be located in the “compensation of employees” section in IM�-compatible data. See also Reserve Bank of India (�005) for a new and smaller number for Indian IT services exports. While important in trade terms, this issue is not so relevant in the employment/revenue terms utilized here, as the fact that most employment/revenues are export-oriented is unaffected.
49. The buyers of 60 percent of the company were two US-based private equity funds, so strictly this transaction would not have shifted ownership between domestic Indian and foreign.
50. See “Merger �ever Breaks Out in Bangalore,” BusinessWeek, April �6, �004,.
51. Looking at NASSCOM’s membership list as off December �1 �005, “Global IT Majors” include (nonexhaustive listing) such household names as Adobe Systems, Ebay, CSC, EDS, Google, HP, IBM, Microsoft, Oracle, SAP, Sun Microsystems, and Yahoo. “Captive BPO Units” include operations for Accenture, Alstrom, AXA, BT, CapGemini, Capital One, Dell, Deloitte, �ord, GE, HSBC, Lehman Brothers, McKinsey, Robert Bosch, Samsung, Siemens, the World Bank, and ThyssenKrupp. Source: NASSCOM (�006b).
��
panies dominate the domestic Indian market, figure 8 indicates that only approximately �� percent of
employment in the Indian IT service sector is with foreign multinationals.5�
Moreover, the link between figure 8’s export revenue and employment shares is not necessarily even
close to one-to-one. Given the reasonably sound assumption that multinationals will be more capital
intensive than local Indian start-up companies, figure 8 probably thus represents a low estimate of the
distribution of India-based employment. However, speculating further about what share of the approxi-
mately �� percent of Indian IT sector employment with foreign multinationals is the result of relocated
jobs and offshoring and offshore outsourcing and what share is from expanding their Indian operations
and integrating them globally is in all probability futile without additional data.
It is nonetheless evident that the tremendous employment success in recent years of the Indian IT
service sector is at least as much and probably more the outcome of domestic entrepreneurs, rather than
foreign multinationals, tapping into the “India advantage” of a large high-skilled, English-proficient talent
pool. The term “domestic entrepreneurs,” though, must be qualified, given the huge influence returnees
from the mostly US-based Indian IT-sector diaspora have had on the growth of the domestic Indian IT
industry.5�
The internationally competitive and heavily export-oriented Indian IT industry emerged in the
1990s in a broad climate of domestic economic liberalization54 and therefore—unlike older sectors of
the Indian economy—was very lightly regulated. The industry is centered predominantly in large Indian
metropolitan areas, most importantly Bangalore, Hyderabad, Chennai, Delhi, and Mumbai. Many IT
company “campuses” in India possess infrastructure (reliable power supply, roads, and telecommunica-
tions) far superior to the surrounding areas and therefore fulfill a role close to a traditional export-oriented
“special economic zone.”55 It is therefore very unlikely that such “campuses” could have thrived outside
these metropolitan areas,56 because this is where the skilled labor they employ exist in large numbers.
In summary, limited relocation of offshoring and offshore outsourcing jobs and dominance of job
creation in metropolitan cities in the Indian IT industry suggest that in developing Asian countries, em-
ployment development paths in the IT service sector, and the service sector more broadly, will be radically
different from those in the manufacturing sector since the 1960s.
5�. This result is the weighted average of the total revenue figures for Indian and multinational companies in figure 8.
5�. In �00� the Indian government acknowledged this influence with the launching of a number of initiatives aimed at the overseas Indian diaspora, especially one primarily for science and technology professionals. See http://stio.nic.in/. It is too early to evaluate the effect of these government initiatives.
54. See Srinivasan (�00�) for an overview of liberalizing reforms since 1991.
55. Such zones were initially proposed in India only as late as �00�. See draft law for special economic zones at http://sezindia.nic.in/draft_central.asp. Their actual introduction in �006 aroused substantial political furor in the country. See “India’s Special Economic Zones,” The Economist, October 1�, �006.
56. It should be noted, though, that with labor costs rising rapidly in places such as Bangalore, the Indian IT industry is starting to spread out into other larger cities. Recall here that India has at least �5 cities with more than 1 million inhabitants.
��
While the broader developmental benefits of IT service industry growth of the scale seen in India in
recent years are self evident—tens of thousands of new high-paying jobs are good for any economy—the
results in this section (and box �) strongly suggest that the IT sector as an export sector alone cannot
replicate the large job gains seen in manufacturing even if it became a large exporter of these new tradable
services. The employment impact of IT services will be similar to that of manufacturing only if the use of
IT services was broadened to other parts of the domestic developing Asian economies.
Domestic use of IT services in developing Asia is all the more required as the most successful
multinational IT service providers from developing Asia—Tata Consulting Systems, Wipro, and Infosys,
to name a few—continue to grow in new markets and may replicate the global business model of their
main OECD competitors. This means that these IT service companies as a result of their own success will
expand their operations and commercial presence outside their home countries. Box � suggests that com-
mercial presence abroad is a substitute rather than a complement in the IT service sector. Hence, success-
ful foreign expansion by domestic multinational IT service companies—for example, Wipro or Infosys
expanding their commercial presence and work forces rapidly in the OECD countries—will present de
veloping Asian home countries with some of the same concerns over the accrual to the domestic economy
in an era of globalization of (no longer) full income from capital and labor usage in the IT sector.
Increased Tradability of Services and Accompanying Broadening Skill Bias
It was demonstrated in section II how technology has facilitated tradability, offshoring and offshore out-
sourcing, and probable elimination of many low-skilled jobs through ongoing automation. While this in-
creases productivity and therefore is welfare enhancing in the long run, the associated skill bias in services
has distributional consequences in that it may restrict some groups’ access to the expanding economic pie.
As the quote from the Indian bank employees’ labor union at the beginning of this working paper
illustrates, this novel tradability may impact hitherto unaffected (protected) service occupations and
sectors in developing Asia’s labor markets in ways similar to what has happened to US IT-related occupa-
tions. This novel tradability can therefore be viewed as a functional equivalent of legislated and negotiated
liberalization through domestic reforms and the World Trade Organization (WTO).5�
A possible decline in the availability of some low-skilled service jobs—or more realistically a less
steep increase in the growth thereof—will pose a challenge for several developing Asian countries, where
very large parts of the labor forces are still employed in agriculture (figure 9). This is particularly so as the
spectacular rise of the Chinese labor-intensive manufacturing sector in recent years has limited the po-
5�. The cited case from the Indian banking sector is a result of principally domestic economic liberalization. Given the suspension of negotiations in �006, the WTO agenda is uncertain.
��
Box 2 Potential impact of the GATS on IT-related services
The General Agreement on Trade in Services (GATS) constitutes a crucial (if unfortunately back-seat) part of the Doha Round of WTO negotiations, which at the time of writing were “in hibernation” and facing an uncertain future. Even though this working paper focuses on the labor market, it is relevant to briefly consider the impact the GATS has already had on trade in services and the impact it could have. At least two of the four modes of service supply defined in the GATS1 (modes � and �) have direct effects on member states’ labor markets.
So far commitments made under the GATS have been significantly deeper and more liberalizing in mode � (commercial presence) than those in mode 1 (cross-border trade).� This situation led Mattoo and Wunsch (�00�, 1�) to conclude that broadly speaking, “the current market opening features of the GATS emulate a foreign investment agreement rather than allowing for trade in the traditional cross-border sense.” Given that traditionally the overwhelming share of services has been regarded as nontradable, deeper liberalization in mode � (and �) through, for instance, opening up for foreign direct investment in more service sectors is preferable when viewed from the perspective of the larger efficiency-enhancing economic impact of the GATS. If countries really wish to improve productivity in their domestic service sectors through the GATS, liberalizing mode � would almost invariably have a far larger effect than liberalizing mode 1.
However, given that IT-related services have in recent years undergone a technology-induced “tradability revolution,” is the above statement still true for this sector in isolation? Certainly, global mode 1 trade in IT-related services has risen dramatically in recent years.� However, has global trade in mode � IT-related services also changed, and if so by how much?
It is a well-established empirical fact in the United States (the only country for which comparable data for total service trade in both modes 1 and � are available for a longer period) that when considering services as a whole, mode � trade, approximated by sales of services through majority-owned foreign affiliates,� has in recent years significantly outstripped traditional mode 1 cross-border exports. In �00� the latest available data show mode � at $��0 billion, fully �0 percent larger than mode 1 trade ($��8 billion) for all services combined. The corresponding “US import side” figures in �00� were $��� billion in mode 1 and $�8� billion in mode �.�
Recreating comparative data for modes 1 and � trade in IT-related services only is highly problematic, given the general dearth of such data and the concerns over respondent confidentiality when countries do collect them. Again, the United States is the only country with relevant data for both modes 1 and �, and even here, constructing a comprehensive time-series is difficult, due to instances of data suppression by the US Bureau of Economic Analysis.
Box figure 1 attempts a comparison of US trade in IT-related services in both modes 1 and �.� The figure compares US mode 1 cross-border trade in “computer and information services” and compares it with mode � trade (sales by US majority-owned foreign affiliates [MOFA] and by foreign majority-owned US affiliates [MOUSA]). Two things stand out from the figure. First, despite the “tradability revolution” in IT-related services, US mode 1 trade in these remains relatively stagnant.� Second, while the data series for mode � is “extremely spotty,” it is clear that the magnitude of US mode � trade in IT-related services is much higher than cross-border mode 1 trade. For US exports in �00�, the ratio is close to �:1, while for US imports the ratio is approximately �.�:1. Compared with total services with a ratio of
1. These are (1) cross-border trade, (�) consumption abroad, (�) commercial presence, and (�) presence of national persons.
�. See WTO (�001). The distinction between modes � and � is frequently somewhat blurred, and hence this statement is true, rela-tive to mode 1, for both modes � and �.
�. According the IMF’s Balance of Payment Statistics (October �00�), total global exports from 108 reporting countries of the more narrowly defined “computer and information services” reached $�� billion in �00�, more than a doubling since �000.
�. Most available data differentiate between “foreign affiliates,” which require more than 10 percent (foreign) ownership, and “majority-owned foreign affiliates,” which require �0 percent or more. As the “nationality implications” of just a 10 percent foreign ownership share are uncertain, all utilized data in this working paper refer to “majority-owned foreign affiliates.” See BEA (1���) for details of referenced US data. Less complete data for other OECD countries are available in OECD (�00�c).
�. Only in 1��� did US mode � sales in services surpass mode 1 exports, while on the import side, this occurred in 1�8�. Source: Bureau of Economic Analysis, table B, www.bea.gov/bea/di/100�serv/tabB.xls (accessed December ��, �00�).
�. Box figure 1 is taken from a presentation by the author at the UNCTAD/ILO/OECD Expert Meeting on ICTs and Development, December �, �00�. All included data are from the US Bureau of Economic Analysis’s most recently available data on trade in services.
�. Note that mode 1 trade in this figure includes both affiliated and unaffiliated trade.(box continues next page)
��
Box 2 (continued)
just 1.�:1, the US IT service sector is thus far more reliant—both on the import and export sides—on mode � service trade.
This has several implications. First, the relative stability of cross-border mode 1 trade in computer and information services when compared with larger and rising mode � trade suggests the two are substitutes rather than complements for—crucially—both US exports and imports.
Second, when mode � remains the dominant channel for both imports and exports in the US IT-related service sector, this again suggests that foreign providers of IT-related services in the US market must continue to expand their commercial presence in the United States, in just the same manner as US IT service firms have done in their foreign markets. This further points to a situation where the home countries of new multinational providers of IT-related services, such as India in the cases of Wipro, Tata Consulting Systems, or Infosys, will not continue to accrue the same share of total income from capital and labor utilized in the IT service industry as before if the heavy focus on exports is maintained.
Third, the dominance of mode � trade in US IT-related service trade suggests that even if the Doha Round is suc-cessfully concluded as it looks now, it may not make much difference for trade in the IT service sector. This is related to both the hitherto paucity of member state offers in the IT service sectors and the exclusion of the “Singapore issues” from the negotiations. Only a renewed effort by WTO members to ensure investment and establishment rights and thus expansion of mode � trade in IT services will have the desired trade-enhancing effects. Breakthroughs in other individual service sectors may improve the prospects for additional benefits in these sectors, too.
Box figure 1 Computer and information services, US modes 1 and 3 trade, 1994–2004 (millions of dollars)
millions of dollars
Note: (D) indicates that data have been suppressed by the BEA for confidentiality reasons. “IT services industry” is defined as sum of “Software publishing,” “Internet service providers, web search portals, data processing services, internet publishing and broadcasting, and other information services” and “Computer system design and related services.”
N/A = not availableMOFAs = majority-owned foreign affiliatesMOUSAs = majority-owned US affiliates
Source: US Bureau of Economic Analysis, Trade in Services, tables at www.bea.gov/bea/di/intiserv.htm.
Box Figure 1: Computer and Information Services, USMode 1 and 3 Trade, 1999-2004 $USmn
10,534 12,253
23,625 24,266
79,124
6643 6822 6723 7093 8453 8939 8239
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
1999 2000 2001 2002 2003 2004 2005
$USm
n
Mode 3: Sales of Services to U.S. Persons by Foreign MNC's in the IT Services Industry Through Their Nonbank MOUSAsMode 3: Sales of Services to Foreign Persons by U.S. MNC's in the IT Services Industry Through Their Nonbank MOFAsMode 1: Total Cross-border Imports of Computer and Information ServicesMode 1: Total Cross-border Exports of Computer and Information Services
(D) (D) (D)(D) (D) (D) (D) (N/A) (N/A)
Note: (D) indicates that data has been suppressed by the BEA for confidentiality reasons. "IT Services Industry" defined as sum of "Software Publishing", "Internet serviceproviders, web search portals, data processing services, internet publishing and broadcasting, and other information services" and "Computer systems design and relatedservices". Source: BEA, Trade in Services tables at http://www.bea.gov/bea/di/intlserv.htm .
��
tential for large-scale moves out of agricultural employment and into low-skilled manufacturing for other
developing Asian countries. �urthermore, accelerating skill bias in service employment will pose a threat
to the substantial unskilled work forces in several developing Asian countries (figure 10).
Any decrease in the employability of developing Asia’s work forces arising from increased tradabil-
ity and automation in services raises the risk that the region may find it increasingly difficult to escape
what has for most East Asian countries been a period of “lower employment elasticity growth” (percent-
age increase in employment associated with a 1 percentage point increase in GDP growth) in the 1990s
compared with earlier periods.58 If this slowdown in East Asian employment growth were to deepen, the
region will find it increasingly difficult to reap the full economic benefits of its growing labor forces in
the coming decade(s), where millions of young people will continue to enter the regional job markets.
�or instance, assuming constant labor force participation rates at �00� levels for the 15-to-64-year-old
population (available in ADB �005) and then utilizing the medium population projections from the UN
World Population Prospects Database (esa.un.org), India faces an increase in its labor force of a staggering
90.� million during the decade �005–15, China faces 56.4 million, Indonesia 15.� million, Vietnam 9
million, the Philippines 8.� million, Thailand � million, and Malaysia �.5 million.
In sum, tradability of services and technological innovation and associated skill bias risks could
curtail the possibilities for developing Asian countries to achieve broad-based and therefore politically
stabilizing employment growth in the future.
Emptying the Competitively Priced Skill Pool Too Early?
Even if the direct employment effects of offshoring and offshore outsourcing have not been dramatic
in overall scope for developing Asian countries, demand especially for skilled professionals and middle
managers has increased. �arrell and Rosenfeld (�005) find that foreign multinationals tend to employ
more than twice the ratio of middle managers to entry-level service agents at offshore locations than in
their home-country operations.59 Unsurprisingly therefore this group of workers has experienced rapid
wage increases in the fastest-growing countries in recent years,60 and most multinationals, as well as local
companies in India, have started complaining about the costs of retaining this crucial group of employ-
58. See ADB (�005), Kapsos (�005), and ILO (�006). See also “Efficiency Made at the Cost of Workers’ Interest,” China Daily, August �0, �006, for estimates that the employment elasticity in China has declined from .45� in the 1980s, to .11 in the 1990s, to .098 in �000–�004.
59. This will partly be a direct result of the need to minimize the risk of miscommunication between units located on different continents, especially if dealing with direct customer interaction.
60. The Hewitt Global Salary Planning Report 2006 estimates that real wage increases in �005 for the group of “Specialists, Professional Staff and Junior Management” rose �.5 percent in Thailand, �.6 percent in Hong Kong, �.� percent in Taiwan, �.� percent in Singapore, �.4 percent in Malaysia, fully 6.� percent in China, and a whopping 11.4 percent in India. Compare these with a corresponding only 0.4 percent in the United States in �005.
�8
ees.61 NASSCOM-McKinsey (�005) even projects a potential 500,000 shortfall in qualified skilled Indian
employees by �010 should current growth rates continue.
Moreover, despite huge headline numbers, several studies indicate that the number of high-skilled
workers in India and elsewhere who will be capable of working for an OECD multinational—and by ex-
tension probably any local company in direct competition with them—is significantly smaller than many
believe.
�or instance, McKinsey Global Institute (�005), based on interviews with multinational human
resource professionals, human resource agencies, and executives, estimates that only 1� percent of low
wage–country university graduates may be both suitable and geographically available for work in a mul-
tinational company.6� Among the Asian countries covered, only 10 percent of Chinese engineers, for in-
stance, are estimated to be potential hires, rising to �0 percent of engineers in the Philippines, �5 percent
in India, and �5 percent in Malaysia. Agarwal (�006) estimates that up to 40 percent of India’s graduates
are unproductively employed and shows that the country indeed has an unemployment rate that rises
with educational attainment.
Meanwhile, OECD (�006b) makes it plain that, in the face of an adverse demographic outlook,
most developed countries are rapidly making it substantially easier for high-skilled workers to immigrate
to OECD countries. As such, despite a tendency for many OECD countries to increasingly restrict overall
levels of immigration, the opposite is the case for high-skilled people.
Kirkegaard (�005b) finds evidence from the US H-1B visa program—which requires that benefiting
entrants into the United States have at least a bachelor’s degree6�—that intensive corporate users of this
program in the US IT sector pay a large group of high-skilled immigrants substantially above prevailing
US wages.64 These companies pay a premium to these foreign entrants to the US labor market, presum-
ably for possessing a unique skill set. This clearly indicates that a global labor market for certain high-
skilled occupations is emerging.
Therefore, it may be stated that not only is the available talent pool in many developing Asian coun-
tries significantly smaller than many would believe based on headline numbers of graduates but also op
61. In the �006 �0-� filings with the Securities and Exchange Commission (SEC), leading listed Indian IT companies—Tata Consulting Systems, Infosys, Wipro, and Satyam—all state rapidly rising wages in India as having an adverse impact on profits and being sufficiently prominent to mention in these filings as a “major risk.” See filings at the SEC Web site, www.sec.gov.
6�. A weighted average of �8 low-wage countries covered in McKinsey Global Institute (�005).
6�. The H-1B program is based on employer sponsorship so that any would-be immigrant for whom the US-based employer is willing to sponsor a visa can enter and work in the United States for up to six years. However, to prevent wage dumping, the program requires that US employers pay the immigrant no less than the prevailing wage for similar workers in the local area. See Kirkegaard (�005b) for an elaboration.
64. However, other studies of US H-1B wages in IT occupations, such as Miano (�005), find to the contrary that wages for H-1B visa holders in IT-related occupations are substantially below prevailing US median wages for these occupations.
��
tions are expanding for many high-skilled people to move to OECD countries to earn much higher wages
than would be possible in their home countries.
�reeman (�006) illustrates this very point from “the other (demand) side,” when he succinctly rebuts
the fears of a looming high-skilled labor shortage in the United States by highlighting the need to look at
the global supply of high-skilled workers:
The lesson from the 1990s increased employment of science and engineering workers is clear: if the US economy demands more highly skilled workers in the period of projected slow labor force growth, it can increase supplies by admitting more immigrants in areas of rising labor demand, as it did in the 1990s. The supply of highly educated persons overseas, many of whom major in science and engineering, suggests that as long as the US is an attractive place to work and is open to immigration, it cannot [emphasis in original] experience a shortage in the science and engineering workforce” (�reeman �006, 15).
Evidently, this has already had a huge positive effect of significant real wage increases for the high-
skilled group in question (see footnote 60) and multinational companies in general pay a nontrivial wage
premium over local employers.65 Subsequently, scarce locally available talent (those not possessing the
desire to become an entrepreneur) can be assumed to principally seek employment at such multinational
companies.
To avoid a scenario where available skilled professionals and middle managers price themselves “out
of the domestic market,” due to their exclusive global opportunities for employment either with multi-
national companies or overseas, developing Asian countries must rapidly raise the quantity and quality
of higher (particularly tertiary) education. Such an improvement should be made predominantly in the
domestic education system. Simply sending more students overseas on scholarships will not produce the
desired impact, as many students on such scholarships may not wish to return home due to expanding
OECD-country employment opportunities.
Yet, as illustrated in the case of the Indian diaspora, whose partial return to India has given a huge
boost to the Indian IT industry, making more talent available at home should never be at the expense of
preventing highly skilled people from leaving. Instead, a dynamic domestic business environment capable
of luring talent—domestic or foreign—to relocate should be created.
In a world economy increasingly characterized by tradability of services and ensuing skill bias in fa-
vor of higher skilled workers, such an environment will prove crucial to generate growth opportunities for
domestic developing Asian service companies capable of competing against global multinationals. Ignor-
ing the Indian IT majors for a minute and looking at the UNCTAD listings of the largest develop
65. See, for instance, Graham (�000) and Lewis and Richardson (�001).
�0
ing-world multinationals,66 it becomes clear that so far no developing country has spawned any significant
multinational company outside the manufacturing sector and incumbent government-run monopolies.
Such potential inability, at least partly due to rapid wage inflation for skilled professionals, of de-
veloping Asian countries to generate internationally successful companies in the most rapidly expanding
sectors of the global economy is in itself troubling. It can unfortunately be expected to provide the basis
for continued calls for protectionism in the service sector—despite the Indian experience that it is domes-
tic companies in open competition with multinationals that provide the largest employment effects (figure
8)—which could have damaging ripple effects on the prospects of global and bilateral trade negotiations.6�
Such calls for protectionism in the face of allegedly superior foreign competitors would regretfully mirror
the behavior of many other special interest groups, most noticeably of course OECD farmers.
Moreover, increasingly unequal access to the “opportunities of the globally integrating economy,”
indicated by a broadening skill bias for service occupations and sectors, and rapidly rising wages for high-
skilled professionals in many countries are increasing income inequality.68 With the majority of high-
skilled service-sector employment likely to occur in metropolitan areas—such as is occurring in India—
heterogeneous developing Asian countries face the additional distributional risk of economic development
becoming more uneven in geographic terms. This lopsided development may aggravate the type of politi-
cal instability seen in, for instance, Thailand, pitting economically advanced capital/metropolitan regions
against underdeveloped outer areas.
IV. SUMMARY AND CONCLUDING REMARKS
This working paper has attempted to gather the available empirical evidence from the main OECD coun-
tries to gauge the extent of production relocation, offshoring, and offshore outsourcing in their labor mar-
kets. The paper develops a three-tier validity hierarchy of available empirical data. Incomplete evidence in
both the United States and EU-15 indicates that this phenomenon has had only a minor effect, with less
than 5 percent of major layoffs in �004–05 attributable to it.
In the EU-15, the majority of layoffs still occur in the manufacturing sector, although 44 percent
now occur in services. Significant differences exist in the intensity of offshoring and offshore outsourcing
among individual EU-15 countries, and these differences are uncorrelated with structural labor-market
policies as well as the level of public anxiety.
66. UNCTAD, World Investment Report 2005, appendix table A.I.10. The OECD Information Technology Outlook 2006, appendix 1, reaches the same inescapable conclusion when looking at ICT multinationals only.
6�. It is a recurring sticking point in US bilateral trade negotiations that developing countries are unwilling to grant US service companies the market access they desire.
68. The United Nations University WIDER database, www.wider.unu.edu/wiid, reports generally widening income inequality in East Asian countries in the late 1990s, reversing earlier trends of low and stable levels of income inequality.
�1
Japanese data show rising overseas employment in manufacturing-sector affiliates in Asia, especially
in China. Survey data indicate that this rise is related mostly to servicing the local market, rather than
the result of production relocation from Japan. Likewise the decline in domestic Japanese manufacturing
employment in all probability is mostly due to accelerating productivity in the domestic Japanese manu-
facturing sector.
Section II surveys different approaches for estimating the share of OECD employment potentially
affected by offshoring and offshore outsourcing especially in the service sectors. Results from different
methodologies indicate that more jobs—roughly between 10 to �0 percent of total employment—may
potentially be affected in the service sectors than in manufacturing in most OECD countries.
�ar from being an irreversible trend of declining employment, however, detailed data on US IT-re-
lated occupations indicate that “being potentially affected by offshoring and offshore outsourcing” rather
implies a rapidly deepening skill bias affecting these types of employment. Low-wage/skilled IT-related
jobs are swiftly disappearing, while high-wage/skilled IT-related jobs continue to be created at a brisk
pace. Evidence is further presented that high-wage/skilled US IT-related occupations were more vulner-
able to rising unemployment during the �001 recession than during earlier cycles, possibly indicating that
this group of workers is entering a novel “cyclical labor market.”
Technological innovation has hitherto facilitated both offshoring and offshore outsourcing, as well as
induced related nontrade skill bias. As the development of new technologies, for instance, voice recogni-
tion software, continues, there is considerable potential for automating many tasks that are today poten-
tial nominees for offshoring and offshore outsourcing. This indicates that the skill level of tasks and jobs
likely to be relocated will also be driven gradually upward, requiring an equal upgrading of destination-
country work force skills if job growth there is to be maintained.
Developing Asian countries will not have seen a dramatic inflow of jobs through offshoring and
offshore outsourcing, and data from the most successful destination country in IT services, India, indicate
that the majority of job creation has been in local Indian companies rather than in foreign multinationals.
This development, and the fact that most jobs in India so far have occurred in metropolitan areas, points
out that employment growth from globalization in the service sectors is likely to be very different from
what developing Asia experienced in the manufacturing sector.
In services, domestic or multinational companies operating in the local market, rather than export-
oriented firms, will create by far the most number of jobs. This means that successful domestic economic
reforms and liberalization—much more than the ability to attract foreign investment for export indus-
tries—will be what guarantee employment growth and economic success. The rise of the service economy
in developing Asia may therefore require quite a dramatic change in focus to the domestic service econo-
my from the earlier emphasis on exports in manufacturing.
��
Increased tradability of services and associated skill bias arising from offshoring and offshore out-
sourcing pose employment creation challenges for those developing Asian countries, where a significant
share of populations are still in agricultural production and possess limited skills. Such population groups
risk facing considerable obstacles in entering rapidly globalizing service sectors and therefore remaining
separated from major vibrant parts of their domestic economies. Thus, this trend may result in increasing
inequality in both economic opportunities and incomes in developing Asia.
Despite limited employment growth resulting directly from offshoring and offshore outsourcing
in developing Asia, key groups of skilled professionals and junior managers in booming countries like
China and India have experienced substantial wage increases in recent years. This highlights the issue
that, despite large headline numbers of high-skilled graduates, only a small group may be able to actually
gain employment at a multinational or similar local company. Deep-pocketed multinational companies
ultimately possess a significant advantage in chasing a scarce resource.
This advantage could—in the absence of significant improvements in the domestic education sys-
tem—hurt the creation, and stunt the growth to multinational status, of domestic service-sector compa-
nies in developing Asia. Such stunted growth of developing Asia’s service-sector companies will not, based
on the Indian experience, be for a lack of entrepreneurial talent, although it is not clear that many other
countries are as well endowed with a huge diaspora as is India. It could nonetheless occur if the supply of
high-skilled talent from the region’s domestic education sector is not increased.
Developing Asian countries therefore face a double educational challenge in the coming years. If
they are to challenge incumbent service-sector multinationals, not only must a large group of the popula-
tions in several countries be taught more than basic numeracy and literacy to be employable outside the
agricultural sector but also many more workers must acquire the advanced skills needed to staff domestic
service companies.
Offshoring and offshore outsourcing, and increased tradability of services and associated skill bias,
will as all trade have an uneven employment impact on developing Asia. Some high-skilled groups are
benefiting and will continue to benefit dramatically from new employment opportunities and rising wage
levels. Meanwhile, the same skill bias may eliminate many employment opportunities for unskilled or
low-skilled groups in the region.
The broader future policy challenge for developing Asia arising from this phenomenon will not
be so much to generate economic growth itself—continued economic liberalization and growing labor
forces will to a large degree ensure that—but to guarantee political stability during economic growth by
ensuring that its benefits and opportunities are distributed reasonably evenly among different groups and
geographic regions. Addressing this challenge will not be easy—many OECD countries, most noticeably
the United States, are struggling with the same challenge—but will be required for long-term growth and
stability.
��
REFERENCES
Agarwal, P. �006. Higher Education in India—The Need for Change. ICRIER Working Paper 180. New Delhi: Indian Council for Research on International Economic Relations.
Amiti, M., and S. Wei. �005. �ear of Service Outsourcing: Is It Justified? Economic Policy 4�: �0�–4�.
Amiti, M., and S. Wei. �006. Services Offshoring and Productivity: Evidence from the United States. NBER Working Paper 119�6 (January). Cambridge, MA: National Bureau of Economic Research.
ADB (Asian Development Bank). �005. Part I—Special Chapter—Labor Markets in Asia: Promoting �ull, Productive, and Decent Employment. Available at www.adb.org.
A.T. Kearney, Inc. �004. Offshore Location Attractiveness Index. Making Offshore Decisions (April).
Aubert, P., and P. Sillard. �005. Delocalisations et reductions d’effectifs dans l’industrie francaise. Serie des Documents de Travail de la Direction des Etudes et Syntheses Economiques. Paris: INSEE.
Autor, D. H., L. �. Katz, and A. B. Krueger. 1998. Computing Inequality: Have Computers Changed the Labor Market? Quarterly Journal of Economics CXIII: 1169–�14.
Autor, D. H., �. Levy, and R. J. Murnane. �00�. The Skill Content of Recent Technological Change: An Empirical Exploration. Quarterly Journal of Economics 118, no. 4 (November).
Bardhan, A., and C. Kroll. �00�. The New Wave of Outsourcing. �isher Center Research Reports. Berkeley, CA: University of California.
BEA (Bureau of Economic Analysis). 1995. A Guide to BEA Statistics on US Multinational Companies. Washington: US Department of Commerce. Available at www.bea.gov (accessed December ��, �006).
Boston Consulting Group. �004. Capturing Global Advantage. BCG Report (April). Boston.
Bhagwati, J., A. Panagariya, and T. N. Srinivasan. �004. The Muddles Over Outsourcing. Journal of Economic Perspectives 18, no. 4 (fall): 9�–114.
Blinder, A. S. �005. Fear of Offshoring. CEPS Working Paper 119 (December). Princeton, NJ: Princeton University.
Blinder, A. S. �006. Activities that Do Not Require Physical Contact or Geographic Proximity Are Most at Risk. CESifo �orum �, no.� (summer). Munich.
Bronfenbrenner, K., and S. Luce. �004. The Changing Nature of Corporate Global Restructuring: The Impact of Production Shifts on Jobs in the US, China, and Around the Globe. Report submitted to the US-China Economic and Security Review Commission, October 14, Washington. Brown, S. P. �004. Mass Layoff Statistics Data in the United States and Domestic and Overseas Relocation. Washington: Bureau of Labor Statistics. Available at www.bls.gov.
��
Desai, Mihir A., C. �ritz �oley, and James R. Hines. �004. �oreign Investment and Domestic Economic Activity. Harvard University. Photocopy.
Desai, Mihir A., C. �ritz �oley, and James R. Hines. �005. Foreign Direct Investment and the Domestic Capital Stock. NBER Working Paper 110�5. Cambridge, MA: National Bureau of Economic Research.
European Commission. �005. Eurobarometer 6�, Spring Wave. Brussels. Available at www.ec.europa.eu.
E�ILWC (European �oundation for the Improvement of Living and Working Conditions). �004. Outsourcing of ICT and Related Services in the EU. Available atwww.emcc.eurofound.eu.int.
�alk, M., and B. M. Koebel. �004. The Impact of Office Machinery and Computer Capital on the Demand for Heterogeneous Labour. Labour Economics 11: 99–11�.
�arrell, D., and J. Rosenfeld. �005. US Offshoring: Rethinking the Response. San �rancisco, CA: McKinsey Global Institute (December).
�eenstra, R. C. 1998. Integration of Trade and Disintegration of Production in the Global Economy. Journal of Economic Perspectives 1�, no. 4: �1–50.
�eldstein, Martin S. 1995. The Effects of Outbound �oreign Direct Investment on the Domestic Capital Stock. In The Effects of Taxation on Multinational Corporations, ed. Martin �eldstein, James R. Hines Jr., and R. Glenn Hubbard. Chicago, IL: University of Chicago Press.
�orrester Research, Inc. �004a. Near-Term Growth of Offshoring Accelerating. Cambridge, MA (May 14).
�orrester Research, Inc. �004b. Two-Speed Europe: Why 1 Million Jobs Will Move Offshore. Cambridge, MA (August 18).
�reeman, Richard B. �006. Labor Market Imbalances: Shortages, or Surpluses, or �ish Stories? Paper presented at the Boston �ederal Reserve Economic Conference on Global Imbalances—As Giants Evolve, Chatham, MA, June 14–16.
Garner, C.A. �004. Offshoring in the Services Sector: Economic Impact and Policy Issues. Federal Reserve Bank of Kansas City Economic Review (Third Quarter).
Gentle, Chris. �004. The Titans Take Hold. Deloitte Research �inancial Services Report. New York: Deloitte Research.
Gerstenberger, B., and R. A. Roehrl . �006. Service Jobs on the Move—Offshore Outsourcing of Business Related Services. In Offshoring and the Internationalization of Employment—A Challenge for a Fair Globalization? Geneva: International Labor Organization and �rench Ministry of Employment.
Global Insight. �004. The Comprehensive Impact of Offshore IT Software and Services Outsourcing on the US Economy and the IT Industry. Lexington, MA (March).
��
GAO (Government Accountability Office). �004. Current Government Data Provide Limited Insight into Offshoring of Services. GAO Report to Congressional Committees GAO-04-9��. Washington.
GAO (Government Accountability Office). �005a. Offshoring of Services—An Overview of the Issues. GAO Report to Congressional Committees GAO-06-5. Washington.
GAO (Government Accountability Office). �005b. U.S. and India Data on Offshoring Show Significant Differences. GAO Report to Congressional Committees GAO-06-116. Washington.
Graham, Edward M. �000. Fighting the Wrong Enemy: Antiglobal Activists and Multinational Enterprises. Washington: Peterson Institute for International Economics.
Industrial Performance Center. �006. Services Offshoring Working Group—Final Report. Cambridge, MA: Massachusetts Institute of Technology.
ILO (International Labor Organization). �006. Labour and Social Trends in Asia and the Pacific �006—Progress Towards Decent Work. Background paper for the 14th ILO Asian Regional Meeting, Busan, Republic of Korea, August �0–September 1. Available at www.ilo.org.
Jensen, J. Bradford, and Lori Kletzer. �005. Tradable Services: Understanding the Scope and Impact of Services Offshoring. Working Paper 05-9. Washington: Peterson Institute for International Economics.
Jensen, P. D. Ø., J. K. Kirkegaard, and N. S. Laugesen. �006. Offshoring in Europe—Evidence of a Two-Way Street from Denmark. Working Paper 06-�. Washington: Peterson Institute for International Economics.
Jorgenson, D.W., and K. Stiroh. �000. Raising the Speed-Limit: US Economic Growth in the Information Age. New York: �ederal Reserve Bank of New York.
Kapsos, S. �005. The Employment Intensity of Growth: Trends and Macroeconomic Determinants. Geneva: Employment Strategy Department, International Labor Organization.
Kimura, �. �005. The Globalization of Production Networks: A View from Asia. Paper presented at the OECD Workshop on “The Globalization of Production: Impacts on Employment, Productivity, and Economic Growth,” November 15–16.
Kirkegaard, Jacob �unk. �004. Outsourcing—Stains on the White Collar? Washington: Peterson Institute for International Economics.
Kirkegaard, Jacob �unk. �005a. Outsourcing and Offshoring: Pushing the European Model Over the Hill Rather Than Off the Cliff! Working Paper 05-1. Washington: Peterson Institute for International Economics.
Kirkegaard, Jacob �unk. �005b. Outsourcing and Skill Imports: Foreign High-Skilled Workers on H-1B and L-1 Visas in the United States. Working Paper 05-15. Washington: Peterson Institute for International Economics.
��
Levy, �., and R. Murnane. �004. The New Division of Labor. New York: Princeton University Press and the Russell Sage �oundation.
Levy, �., and R. Murnane. �006a. How Computerized Work and Globalization Shape Human Skill Demands. Cambridge, MA: Massachusetts Institute of Technology. Available at http://web.mit.edu.
Levy, �., and R. Murnane. �006b. What Sort of Rich-Country Jobs (and How Many) Are at Risk to Globalization? CESifo Forum �, no. � (summer). Munich.
Lewis III, Howard, and J. David Richardson. �001. Why Globalization Really Matters. Washington: Peterson Institute for International Economics.
Mankiw, N. G., and P. Swagel. �005. The Politics and Economics of Offshore Outsourcing. Washington: American Enterprise Institute (November).
Mann, Catherine, with Jacob �unk Kirkegaard. �006. Accelerating the Globalization of America: The Role for Information Technology. Washington: Peterson Institute for International Economics.
Mattoo, Aaditya, and Sacha Wunsch Vincent. �004. Preempting Protectionism in Services: The GATS and Outsourcing. Washington: Peterson Institute for International Economics.
McKinsey Global Institute. �005. The Emerging Global Labor Market: Part II—The Supply of Offshore Talent in Services. San �rancisco, CA: McKinsey and Company.
Miano, John. �005. The Bottom of the Pay Scale—Wages for H-1B Computer Programmers. Center for Immigration Studies Backgrounder. Washington: Center for Immigration Studies.
METI (Ministry of Economy, Trade and Industry, Japan). �004. White Paper on International Economy and Trade 2004. Tokyo.
METI (Ministry of Economy, Trade and Industry, Japan). �006. White Paper on International Economy and Trade 2006. Tokyo.
NASSCOM-McKinsey. �005. NASSCOM-McKinsey Study 2005. Available for a fee at www.nasscom.in.
NASSCOM (National Association of Software and Service Companies). �006a. The Indian IT Industry—Fact Sheet. Available at www.nasscom.in.
NASSCOM (National Association of Software and Service Companies). �006b. NASSCOM Annual Report 2006. Available at www.nasscom.in.
OECD (Organization for Economic Cooperation and Development). �004. Employment Outlook 2004. Paris.
OECD (Organization for Economic Cooperation and Development). �005a. Handbook on Economic Globalization Indicators. Paris.
��
OECD (Organization for Economic Cooperation and Development). �005b. OECD Compendium of Productivity Indicators 2005. Paris.
OECD (Organization for Economic Cooperation and Development). �005c. OECD Statisticson Measuring Globalization, Volume II: Services (�ATS). Paris.
OECD (Organization for Economic Cooperation and Development). �006. Trends and Recent Developments in Foreign Direct Investment (June). Paris.
OECD (Organization for Economic Cooperation and Development). �006b. International Migration Outlook 2006. Paris.
OECD (Organization for Economic Cooperation and Development). �006c. OECD Information Technology Outlook. Paris.
Reserve Bank of India. �005. Computer Services Exports from India: �00�–0�. RBI Monthly Bulletin (September). Mumbai, India.
Roland Berger and UNCTAD. �004. Service Offshoring Takes Off in Europe—In Search of Improved Competitiveness. Geneva (June).
Schulze, C. L. �004. Offshoring, Import Substitution and the Jobless Recovery. Brookings Institution Policy Brief 1�6 (August). Washington: Brookings Institution.
Slaughter, Matthew J. �004. Insourcing Jobs: Making the Global Economy Work �or America. Organization for International Investment. Photocopy (October).
Srinivasan, T. N. �00�. Indian Economic Reforms: A Stocktaking. New Haven, CT: Economic Growth Center, Yale University.
UNCTAD (UN Conference on Trade and Development). �004. World Investment Report 2004. Geneva.
UNCTAD (UN Conference on Trade and Development). �005. World Investment Report 2005. Geneva.
UNESCO/OECD. �005. Education Trends In Perspective: Analysis of the World Education Indicators—2005 Edition. Paris.
Van Ark, B. �005. Does the European Union Need to Revive Productivity Growth? Working Paper GD-�5. Groningen, the Netherlands: Groningen Growth and Development Center.
Van Welsum, D., and G. Vickery. �005. Potential Offshoring of ICT-Intensive using Occupations. OECD Working Party on the Information Economy Report DSTI/ICCP/IE(�004)19/�inal. Paris: Organization for Economic Cooperation and Development.
Van Welsum, D., and G. Vickery. �006. The Share of Employment Potentially Affected By Offshoring—An Empirical Investigation. OECD Working Party on the Information Economy Report DSTI/ICCP/IE(�005)8/�inal. Paris: Organization for Economic Cooperation and Development.
�8
WTO (World Trade Organization). �001. Market Access: Unfinished Business—Post–Uruguay Round Inventory and Issues. Geneva: WTO Economic Research and Analysis Division.
Zeile, W. J. �00�. Trade in Goods Within Multinational Companies: Survey-Based Data and �indings for the United States of America. Bureau of Economic Analysis paper prepared for the OECD Committee on Industry and Business Environment Working Party on Statistics, Session on Globalization, Paris, November �–4.
Zeile, W. J. �005. U.S. Affiliates of �oreign Companies: Operations in �00�. Survey of Current Business (August). Washington: Bureau of Economic Analysis, US Department of Commerce.
��
Table 1 US job losses from mass layoffs associated with domestic and overseas production relocation, 2004Q1–2005Q4
1 Total number of separations, excluding seasonal and vacation related
���,0�8
2 Total number of separations associated with movement of work 11�,�0�
3 Percent share of total separations 11.�
4 Total number of separations associated with movement of work for which employers were able to provide specific information regarding the movement of work component (�� percent of 11�,�0�)
8�,�8�
5
Percent share of separations associated with movement of work domestically and within company (domestic inhouse relocation)
��.� (�.�)
6
Percent share of separations associated with movement of work domestically and to a nonaffiliated producer
11.� (1.�)
7
Percent share of separations associated with movement of work internationally and within company (offshoring)
��.� (�.1)
8
Percent share of separations associated with movement of work internationally and to a nonaffiliated producer (offshore outsourcing)
�.� (0.8)
100.0 (11.�)
Note: Share of total separations (���, 0�8) in parentheses.
Sources: Brown (�00�) and quarterly BLS MLS news releases, www.bls.gov.
�0
Table 2 Total EU-15 job losses, by type of restructuring, 2004-05
Type of restructuring (ERM definition)
Share of total jobs lost/planned
reductions (percent)
Outsourcing (when the activity is subcontracted to another company within the same country) 0.�
Relocation (when the activity stays within the same company but is relocated to another location within the same country) 1.�
Merger/acquisition (when two companies merge or during an acquisitionwhich then involves an internal restructuring program aimed at rationalizing organization by cutting personnel) �.�
Offshoring/delocalization (when the activity is relocated or outsourced outside of the country’s borders) �.�
Bankruptcy/closure (when an industrial site is closed or a company goesbankrupt for economic reasons not directly connected to relocation oroutsourcing) 1�.�
Internal restructuring (when the company undertakes a job-cutting plan, which is not linked to another type of restructuring defined above) ��.0
Other 0.1
Total (1,08�,��� job losses included) 100.0
Source: European Restructuring Monitor, www.emcc.eurofound.eu.int.
�1
Table 3 Detailed US IT-related occupations, 1999 and May 2005 (number of workers)
Occupation 1999 May 2005 Total changePercentage
change
Annual wage,May 2005
(dollars)
Annual real wage change,
1999–May 2005 (percent)
Call center type occupations Telemarketers �8�,��0 �00,8�0 –8�,��0 –1�.� ��,�00 –1.1 Telephone operators �0,8�0 ��,��0 –�1,��0 –��.� �1,0�0 –0.�Low-wage technology workers Switchboard operators, including answering service ��8,��0 1��,�80 –��,��0 –�1.� ��,0�0 –0.1 Computer operators 1�8,�00 1��,1�0 –��,��0 –��.� ��,�80 0.� Data entry keyers ��0,��0 ���,�00 –���,��0 –��.0 ��,�10 0.� Word processors and typists ��1,�10 1��,�80 –11�,��0 –��.� �0,1�0 1.� Desktop publishers ��,0�0 ��,�10 –�,1�0 –1�.� ��,��0 –1.0 Electrical and electronic equipment assemblers �8�,��0 �0�,��0 –180,1�0 –��.� ��,1�0 1.� Semiconductor processors ��,110 ��,��0 �,�10 �.� ��,8�0 0.�Total call center and low-wage technology workers 2,241,650 1,486,470 –755,180 –33.7 26,446 0.�
Mid-level IT workers Computer support specialists ���,8�0 ���,8�0 ��,0�0 8.0 ��,�80 –1.� High-wage technology workers Computer and information scientists, research ��,�80 ��,8�0 –��0 –1.� ��,0�0 �.� Computer programmers 528,600 389,090 –139,510 –26.4 67,400 1.1 Computer software engineers, applications �8�,�00 ���,�80 1�8,�80 �8.� ��,��0 0.8 Computer software engineers, systems software �0�,0�0 ��0,��0 111,��0 ��.� 8�,�10 �.0 Computer systems analysts ��8,�10 ���,1�0 ��,�10 1�.� �0,��0 1.0 Database administrators 101,��0 ��,�80 –�,080 –�.1 ��,��0 1.� Network and computer systems administrators �0�,�80 ��0,��0 ��,��0 ��.1 ��,�10 1.8 Network systems and data communications analysts �8,��0 18�,1�0 8�,8�0 88.� ��,��0 0.0 Computer hardware engineers �0,��0 �8,�80 18,1�0 �0.1 8�,1�0 �.� Electrical engineers 1��,�10 1��,��0 –�,��0 –�.� ��,0�0 1.� Electronics engineers, except computer 10�,8�0 1�0,0�0 ��,��0 �1.� ��,��0 1.8Total high-wage technology workers 2,200,650 2,592,250 391,600 17.8 73,504 1.�
Source: Bureau of Labor Statistics CES data, 1���, �000, �001, �00�, May �00�, November �00�, May �00�, November �00�, and May �00� National Occupational Employment and Wage Estimates.
��
Ownership
LocationInternal (inhouse) External
Domestic
Domestic
internal
production
Domestic external production
by nonaffiliated producer
(outsourced production)
In foreign country
(cross-border/offshore)
Production within group (inhouse)
in foreign country
(offshored production)
Production outside group
in foreign country
by nonaffiliated producer
(offshore outsourced production)
Offsh
orin
g
Offshore outsourcingOutsourcing
Figure 1 Standard offshoring and outsourcing matrix for production relocation
Source: Adapted from UNCTAD (�00�, table IV.1).
��
Location
Domestic(movement within a country)
Foreign(movement between countries)
Inhouse Nonaffiliated Inhouse Nonaffiliated
Type of
sourcing
OutsourcingTo other
inhouse domestic production location
Domestic outsourcing: To other domestic
nonaffiliated production location
Offshoring:To foreign inhouse
production location
Offshore outsourcing: To foreign
nonaffiliated production location
Insourcing
From otherinhouse domestic
production location
Domestic insourcing: From other domestic
nonaffiliated production location
Inshoring:From foreign
inhouse production location
Offshore insourcing: From foreignnonaffiliated
production location
Note: Shaded cells also part of standard matrix in figure 1.
Source: Adapted from Van Welsum and Vickery (�00�) and UNCTAD (�00�).
Figure 2 Complete production relocation options matrix
��
F ig u r e 3 O ffs h o r in g /o ffs h o r e o u ts o u r c in g in te n s ity m a jo r la y o ffs /to ta l p r iv a te e m p lo y m e n t, 2 0 0 4 -0 5
��1
�1�
�1��1�
1��
1��
118
�0��
�� �0�� ��
18
0 0
100
0
�0
100
1�0
�00
��0
�00
��0
�00
��0
Finland Portugal Britain Denmark Ireland Sweden France EU-1� Euro-zone Belgium Netherlands Austria Germany Italy Spain Greece Luxembourg
E U -1 5 = 1 0 0
Source: European Restructuring Monitor ; Eurostat; author's calculations.
Note: Private employment defined as total employment minus NACE category L (public administration and defense; compulsory social security).
Figure 3 Offshoring/offshore outsourcing intensity, major layoffs/total private employment, 2004–05
index
Source: European Restructuring Monitor; Eurostat Labor Force Survey Statistics; author’s calculations.
��
Greece
Spain
Italy
Germany
Austria
Netherlands
Belgium
France
Ireland
Britain
Portugal
Finland
Greece
Spain
Italy
Germany
Austria
Netherlands
Belgium
France
Sweden
Ireland
Denmark
Britain
Portugal
Finland
y = -0.0018x + 3.6299R2 = 0.0727
y = 3E-05x + 0.3936R2 = 0.0006
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
0 50 100 150 200 250 300 350 400
offshoring intensity (EU-15 = 100)
OE
CD
EP
L in
dex,
200
3 (h
ighe
r =
mor
e re
stric
tive)
10%
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
publ
ic fe
ar o
f offs
horin
g/of
fsho
re o
utso
urci
ng 1
)
Public Fears EPL Legislation
Note: Share of people that perceived the word "Globalization" foremost linked to the statement "Delocalization of some companies to countries where laboour is cheaper". Source: European Restructuring Monitor; European Commission - Eurobarometer 63 (2005); OECD (2004); author's calculations.
Figure 4 EU-15 intensity, public fears, and collective dismissal EPL legislation
OEC
D/E
PL in
dex
�00�
(hig
her =
mor
e re
stric
tive)
public fear of offshoring/offshore outsourcing (percent)
EPL = employment protection legislation
Note: Right axis shows percent of people who perceived the word “globalization” foremost linked to “delocalization of some companies to countries where labor is cheaper.”
Sources: European Restructuring Monitor; European Commission (�00�); OECD (�00�); author’s calculations.
��
millions of employees at affiliatesmillions of employees in Japan
F ig u r e 5 E m p lo y m e n t in J a p a n 's m a n u fa c tu r in g s e c to r a n d a t J a p a n e s e -o w n e d m a n u fa c tu r in g fir m s , 2 0 0 1 -0 6
11.�
1�.8 �.8
1.�
11
11.�
11.�
11.�
11.8
1�
1�.�
1�.�
1�.�
1�.8
1�
�001Q�
�001Q�
�001Q�
�00�Q1
�00�Q�
�00�Q�
�00�Q�
�00�Q1
�00�Q�
�00�Q�
�00�Q�
�00�Q1
�00�Q�
�00�Q�
�00�Q�
�00�Q1
�00�Q�
�00�Q�
�00�Q�
�00�Q1
Mill
ions
of E
mpl
oyee
s in
Jap
an
1.�
1.�
1.�
�.1
�.�
�.�
�.�
�.�
Mill
ions
of E
mpl
oyee
s A
t Affi
liate
s
Japanese Manufacturing Employment
Employment at Japanese Overseas Manufacturing Affiliates
Source: Japanese Labor Force Survey and METI Survey of Trends in Business Activities of Japanese Foreign Affiliates and since �00� Quarterly Survey of Overseas Subsidiaries (Kaiji Chosa).
Figure 5 Employment in Japan’s manufacturing sector and at Japanese-owned manufacturing firms, 2001-06
Sources: Japanese Labor Force Survey; METI Survey of Trends in Business Activities of Japanese Foreign Affiliates; METI Quarterly Survey Since �00� of Overseas Subsidiaries (Kaiji Chosa).
��
Figure 6 Japanese overseas manufacturing affiliate employment, by region
ASEAN-� = Malaysia, Philippines, Indonesia, and ThailandNIEs-� = Singapore, South Korea, and Taiwan
Note: The survey methodology was altered beginning �00�.
Sources: METI Survey of Trends in Business Activities of Japanese Foreign Affiliates; METI Quarterly Survey Since �00� of Overseas Subsidiaries (Kaiji Chosa).
Figure 4: Japanese Overseas Manufacturing Affiliate Employment, By Region
0
500000
1000000
1500000
2000000
2500000
3000000
2001Q2 2001Q3 2001Q4 2002Q1 2002Q2 2002Q3 2002Q4 2003Q1 2003Q2 2003Q3 2003Q4 2004Q1 2004Q2 2004Q3 2004Q4 2005Q1 2005Q2 2005Q3 2005Q4 2006Q1
# of
Em
ploy
ees
China (incl. Hong Kong) ASEAN-4 NIES-3 Other Asia Europe North America Other
Note: The survey methodology was altered beginning 2003. ASEAN-4 = Malaysia, Philippines, Indonesia and Thailand. NIES-3 = Singapore, South Korea and Taiwan.Source: METI Survey of Trends in Business Activities of Japanese Foreign Affiliates and since 2003 Quarterly Survey of Overseas Subsidiaries (Kaiji Chosa)
number of employees
�8
Figure 7 US total unemployment rate and select categories of IT-related occupations
0
2
4
6
8
10
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
Q4 2000
Q1 2001
Q2 2001
Q3 2001
Q4 2001
Q1 2002
Q2 2002
Q3 2002
Q4 2002
Q1 2003
Q2 2003
Q3 2003
Q4 2003
Q1 2004
Q2 2004
Q3 2004
Q4 2004
Q1 2005
Q2 2005
Q3 2005
Q4 2005
Q1 2006
Computer Scientists and System Analysts Computer Programmers Computer Software Engineers
Computer Support Specialists Computer Hardware Engineers Electrical and Electronic Engineers
Total US Unemployment Rate
Note: Annual rate from 1983 to 1999 and 4-quarter moving averages from 2000-2006.Sources: Economic Policy Institute ; US Bureau of Labor Statistics.
NBER recession period
Note: Annual rate from 1��� to 1��� and �-quarter moving averages from �000 to �00�.
Sources: Economic Policy Institute; US Bureau of Labor Statistics.
Figure 7 US total unemployment rate and select categories of IT-related occupations
percent
��
Figure 8 India-based IT service providers, export revenue by corporate characteristics, fiscal 2005 (billions of dollars)
4.5
2.5
1.5 1.5
0.2 0.2
0.9
0.2
0.7
2.3
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Top-4 Indian ITCompanies
Tier II Indian ITCompanies
Indian BPO-OnlyCompanies
Upstart IndianCompanies
Offshore Operations ofGlobal IT Majors
Captive BPO Units (Non-IT Multinationals)
IT Enabled Services (ITES) Busines Process Offshoring (BPO)
Indian companies total share: ITES: 85 percent BPO: 35 percent
Foreign multinationals' total share: ITES: 15 percent BPO: 65 percent
Source: NASSCOM (�00�a).
IT enabled services (ITES) Business process outsourcing (BPO)
�0
Source: Asian Development Bank, Development Indicators, 2005: Labor Markets in Asia, available at www.adb.org.
Figure 9 Employment by gender and sector, 2002
percent
Figure 7: Employment by Gender and Sector 2002
0.1 0.2
11.6 1424.9
43.147.5
0.3 0.49.3
20.6
45.4 43.349.8
0%
20%
40%
60%
80%
100%
Hong K
ong
Singap
oreSou
th Kore
a
Malays
ia
Philipp
ines
Indon
esia
Thaila
nd
Hong K
ong
Singap
oreSou
th Kore
a
Malays
ia
Philipp
ines
Indon
esia
Thaila
ndShare of Workforce in Agriculture Share of Workforce in Manufacturing Share of Workforce in Services
Female employment Male employment
Source: ADB Key Indicators 2005: Labor Markets in Asia
100
80
�0
�0
�0
0
�1
Figure 10 Educational attainment of adult population 25–64 years of age, 2002–03
Note: For statistical purposes and its small size, the category “At least primary” is included in “Lower secondary” for Japan, Australia, and Britain.
Source: UNESCO/OECD (�00�, Annex table 1.1).
Figure 8: Educational Attainment of Adult Population 25-64years of Age 2002/2003
0 0 0 3 3 5 614 14 16 20 25 30 34 37 37 42 43 44 44
53 56 5864 64 68 68
6369 72
42
7357 51
57 61 61
69 6244
4246
5249 42 42 38
3135
38 25 27 23 1937
31 28
55
24
38 43
29 25 2311 13
26 2417
11 1015 14 18 16
9 411 9 9 12
0
10
20
30
40
50
60
70
80
90
100
Japa
nAus
tralia
Brita
inRus
siaGerm
any
United
States
Can
ada
Korea
OECD mea
n Fran
ce
Italy
Chile
Spain
Jorda
n Gree
ce
Malays
iaUrug
uay
Non-O
ECD Mea
nArge
ntina
PeruMex
icoBraz
il Ind
ones
iaPort
ugal
Turkey
Para
guay
Thaila
nd
Shar
e of
Tot
al 2
5-64
y Po
pula
tion
At least Primary (Unskilled) At Least Lower Secondary (Skilled) At Least Some Tertiary
Source: UNESCO/OECD (2005), Annex table 1.1.
Note: For statistical gathering purposes and its small size, the category "At Least Primary" is included in "Lower Secondary" for Japan, Australia and Britain
Non-O
ECD mea
n
percent
top related